FAQ
How many distinct values are there of (p_log_file_date, p_log_file_min) in
the source table?
e.g. select p_log_file_date, p_log_file_min, count(*)
from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

Hi

My sql is simple like this:
Insert overwrite table t1 select xxxx from t2

t2 is a table with 5 1GB source file. After insert query, impala generated
one single 4GB file.
There are 20 impala-server nodes in my env.

I donnot know how to make the output distributed which will make the query
much faster.
Can anybody help?

Search Discussions

  • 邱睿 at Aug 9, 2013 at 2:54 am
    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min) in
    the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*)
    from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;



    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 <rayqi...@gmail.com <javascript:>>wrote:
    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • 邱睿 at Aug 9, 2013 at 4:21 am
    So can you tell the way to use RANDOM?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same
    value so a single file is produced.

    Do you have table & column statistics on the source table? If not, gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which
    should give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min




    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 <rayqi...@gmail.com <javascript:>>wrote:
    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*)
    from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • 邱睿 at Aug 9, 2013 at 5:25 am
    I analyzed the source table. After than insert query failed with exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in
    position 138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
       PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same
    value so a single file is produced.

    Do you have table & column statistics on the source table? If not, gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which
    should give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min




    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 <rayqi...@gmail.com <javascript:>>wrote:
    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*)
    from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • 邱睿 at Aug 9, 2013 at 5:25 am
    I analyzed the source table. After than insert query failed with exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in
    position 138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
       PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same
    value so a single file is produced.

    Do you have table & column statistics on the source table? If not, gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which
    should give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min




    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 <rayqi...@gmail.com <javascript:>>wrote:
    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*)
    from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • Greg Rahn at Aug 9, 2013 at 7:49 pm
    Currently there is not a way to hint the distribution for the write/insert.


    On Thu, Aug 8, 2013 at 10:25 PM, 邱睿 wrote:


    I analyzed the source table. After than insert query failed with exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in
    position 138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
    PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same
    value so a single file is produced.

    Do you have table & column statistics on the source table? If not,
    gather them and see if that changes it from HASH_PARTITIONED to RANDOM
    which should give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_**partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min



    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 wrote:


    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,**Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*)
    from qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • Alex Behm at Aug 9, 2013 at 8:29 pm
    Hi,

    some more questions to clarify the issue:
    1. Is the source table a partitioned or unpartitioned table?
    2. Can you provide the CREATE TABLE of the source table (t2)?
    3. Can you provide the full query plan of the plan with PLAN FRAGMENT
    0 being UNPARTITIONED?

    Thanks!

    Alex
    On Thu, Aug 8, 2013 at 10:25 PM, 邱睿 wrote:

    I analyzed the source table. After than insert query failed with exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in position
    138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
    PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same value
    so a single file is produced.

    Do you have table & column statistics on the source table? If not, gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which should
    give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min



    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 wrote:


    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*) from
    qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • Alex Behm at Aug 10, 2013 at 12:22 am
    One more suggestion:
    Since you are trying to insert into a single partition, it'd be better
    to use a static partition insert rather than a dynamic partition
    insert. To clarify:

    Dynamic partition insert (the original query you pasted):
    insert OVERWRITE TABLE qb__s_dm_date_partition_log PARTITION
    (p_event_date, p_event_min) ...

    Static partition insert (try using this instead):
    insert OVERWRITE TABLE qb__s_dm_date_partition_log PARTITION
    (p_event_date='2013-08-09', p_event_min='0935') ...

    With the static partition insert I'm pretty sure Impala will not hash
    repartition because Impala can easily determine that you are inserting
    into a single partition.
    (With the dynamic partition insert it's rather tricky to figure out
    that you are inserting into a single partition, i.e., we'd have to
    apply the value constraints from the WHERE clause to the columns in
    the PARTITION clause).

    Hope it works!

    Cheers,

    Alex
    On Fri, Aug 9, 2013 at 1:29 PM, Alex Behm wrote:
    Hi,

    some more questions to clarify the issue:
    1. Is the source table a partitioned or unpartitioned table?
    2. Can you provide the CREATE TABLE of the source table (t2)?
    3. Can you provide the full query plan of the plan with PLAN FRAGMENT
    0 being UNPARTITIONED?

    Thanks!

    Alex
    On Thu, Aug 8, 2013 at 10:25 PM, 邱睿 wrote:

    I analyzed the source table. After than insert query failed with exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in position
    138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
    PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same value
    so a single file is produced.

    Do you have table & column statistics on the source table? If not, gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which should
    give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min



    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 wrote:


    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date, p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*) from
    qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make the
    query much faster.
    Can anybody help?
  • 邱睿 at Aug 12, 2013 at 7:38 am
    Thanks Alex, the static partition insert works. The plan:

    ----------------
    PLAN FRAGMENT 0
       PARTITION: RANDOM

       WRITE TO HDFS table=default.qb__s_dm_date_partition_log
         overwrite=true
         partitions: '2013-08-12','1445'

       0:SCAN HDFS
          table=default.qb__s_dm_base_log #partitions=1 size=9.04GB
          predicates: length(app_key) = 12, app_key != 'MUNTOAINGVJ8', app_key != 'AXHPMT93XF3Q'
          tuple ids: 0
    ----------------


    But planning time cost more than 1mins. I'not know if there're any way to improve:

         Query Timeline: 2m29s
            - Start execution: 1.709ms (1.709ms)
            - Planning finished: 1m57s (1m57s)
            - Request finished: 2m29s (32s807ms)
            - Unregister query: 2m29s (39.359ms)




    在 2013年8月10日星期六UTC+8上午8时21分57秒,Alex Behm写道:
    One more suggestion:
    Since you are trying to insert into a single partition, it'd be better
    to use a static partition insert rather than a dynamic partition
    insert. To clarify:

    Dynamic partition insert (the original query you pasted):
    insert OVERWRITE TABLE qb__s_dm_date_partition_log PARTITION
    (p_event_date, p_event_min) ...

    Static partition insert (try using this instead):
    insert OVERWRITE TABLE qb__s_dm_date_partition_log PARTITION
    (p_event_date='2013-08-09', p_event_min='0935') ...

    With the static partition insert I'm pretty sure Impala will not hash
    repartition because Impala can easily determine that you are inserting
    into a single partition.
    (With the dynamic partition insert it's rather tricky to figure out
    that you are inserting into a single partition, i.e., we'd have to
    apply the value constraints from the WHERE clause to the columns in
    the PARTITION clause).

    Hope it works!

    Cheers,

    Alex
    On Fri, Aug 9, 2013 at 1:29 PM, Alex Behm wrote:
    Hi,

    some more questions to clarify the issue:
    1. Is the source table a partitioned or unpartitioned table?
    2. Can you provide the CREATE TABLE of the source table (t2)?
    3. Can you provide the full query plan of the plan with PLAN FRAGMENT
    0 being UNPARTITIONED?

    Thanks!

    Alex

    On Thu, Aug 8, 2013 at 10:25 PM, 邱睿 <rayqi...@gmail.com <javascript:>>
    wrote:
    I analyzed the source table. After than insert query failed with
    exception:
    UnicodeEncodeError: 'ascii' codec can't encode character u'\xdc' in
    position
    138871: ordinal not in range(128)

    I found a bug on jira similar to this error. I think they're cause by
    the
    same reason.

    Besides, the query plan shows unparititoned:

    PLAN FRAGMENT 0
    PARTITION: UNPARTITIONED


    Can I choose random partition in maunal?

    在 2013年8月9日星期五UTC+8上午11时03分41秒,Greg Rahn写道:
    No, the file count is based on the write distribution type. In this
    query, the write distribution is highlighted below, so with only a
    single
    value of (p_log_file_date, p_log_file_min), all rows hash to the same
    value
    so a single file is produced.

    Do you have table & column statistics on the source table? If not,
    gather
    them and see if that changes it from HASH_PARTITIONED to RANDOM which
    should
    give you one file per DataNode.

    ----------------
    PLAN FRAGMENT 0
    PARTITION: HASH_PARTITIONED: p_log_file_date, p_log_file_min

    WRITE TO HDFS table=default.qb__s_dm_date_partition_log
    overwrite=true
    partitions: p_log_file_date,p_log_file_min



    On Thu, Aug 8, 2013 at 7:54 PM, 邱睿 wrote:


    only one.

    Is impala output file num consistent to this partition count?

    在 2013年8月9日星期五UTC+8上午10时36分54秒,Greg Rahn写道:
    How many distinct values are there of (p_log_file_date,
    p_log_file_min)
    in the source table?
    e.g. select p_log_file_date, p_log_file_min, count(*) from
    qb__s_dm_base_log group by p_log_file_date, p_log_file_min;


    On Thu, Aug 8, 2013 at 7:00 PM, 邱睿 wrote:

    Hi

    My sql is simple like this:
    Insert overwrite table t1 select xxxx from t2

    t2 is a table with 5 1GB source file. After insert query, impala
    generated one single 4GB file.
    There are 20 impala-server nodes in my env.

    I donnot know how to make the output distributed which will make
    the
    query much faster.
    Can anybody help?
  • 邱睿 at Aug 12, 2013 at 7:39 am
    Here're the complete query profile


    Query (id=c7406bd47560e48d:e6576df75f99ef99):
       Summary:
         Start Time: 2013-08-12 14:47:17.211323000
         End Time: 2013-08-12 14:49:47.169681000
         Query Type: DML
         Query State: FINISHED
         Query Status: OK
         Impala Version: impalad version 1.1 RELEASE (build 5e15fcacc48ec4ea65e8aa76362cb3ec9be26f13)
         User: ude
         Default Db: default
         Sql Statement: insert OVERWRITE TABLE qb__s_dm_date_partition_log PARTITION(p_event_date='2013-08-12', p_event_min='1445')
    SELECT sdk_version,
            log_type,
            app_key,
            device_id,
            android_id,
            random_num,
            resolution,
            session_id,
            timestamp_first,
            timestamp_pre,
            timestamp_cur,
            event_date,
            session_visits,
            event_category,
            event_action,
            event_label,
            event_value,
            param1_key,
            param1_value,
            param1_scope,
            param2_key,
            param2_value,
            param2_scope,
            param3_key,
            param3_value,
            param3_scope,
            param4_key,
            param4_value,
            param4_scope,
            param5_key,
            param5_value,
            param5_scope,
            app_version_name,
            app_version_code,
            os_version,
            language,
            country,
            device_model,
            leos_version,
            manufacturer,
            client_ip,
            network_type,
            lps_did_type,
            lps_did,
            server_timestamp,
            app_channel,
            date_year,
            date_quarter,
            date_month,
            date_week,
            date_day,
            date_hour,
            country_code,
            region_code,
            city,
            latitude,
            longitude,
            operators,
            refer_path,
            preserve1,
            preserve2,
            if(preserve2!='','1','0'),
            preserve4,
            preserve5,
            preserve6,
            preserve7,
            preserve8,
            preserve9,
            preserve10
    FROM qb__s_dm_base_log
    WHERE p_log_file_date='2013-08-12'
       AND p_log_file_min='1445'
       AND length(app_key) = 12
       AND app_key != 'MUNTOAINGVJ8'
       AND app_key != 'AXHPMT93XF3Q'
         Plan:
    ----------------
    PLAN FRAGMENT 0
       PARTITION: RANDOM

       WRITE TO HDFS table=default.qb__s_dm_date_partition_log
         overwrite=true
         partitions: '2013-08-12','1445'

       0:SCAN HDFS
          table=default.qb__s_dm_base_log #partitions=1 size=9.04GB
          predicates: length(app_key) = 12, app_key != 'MUNTOAINGVJ8', app_key != 'AXHPMT93XF3Q'
          tuple ids: 0
    ----------------
         Query Timeline: 2m29s
            - Start execution: 1.709ms (1.709ms)
            - Planning finished: 1m57s (1m57s)
            - Request finished: 2m29s (32s807ms)
            - Unregister query: 2m29s (39.359ms)
       ImpalaServer:
          - ClientFetchWaitTimer: 0ns
          - RowMaterializationTimer: 0ns
       Execution Profile c7406bd47560e48d:e6576df75f99ef99:(Active: 21s882ms, % non-child: 0.00%)
         Per Node Peak Memory Usage: T-HD-10:22000(561.36 MB) T-HD-23:22000(592.74 MB) T-HD-22:22000(561.36 MB) T-HD-06:22000(552.36 MB) T-HD-09:22000(551.36 MB) T-HD-24:22000(553.86 MB) T-HD-15:22000(544.86 MB) T-HD-11:22000(472.98 MB) T-HD-20:22000(520.36 MB) T-HD-19:22000(480.98 MB) T-HD-16:22000(569.36 MB) T-HD-03:22000(410.25 MB) T-HD-18:22000(541.48 MB) T-HD-17:22000(552.86 MB) T-HD-02:22000(394.25 MB) T-HD-13:22000(582.48 MB) T-HD-14:22000(642.26 MB) T-HD-21:22000(504.48 MB) T-HD-07:22000(616.75 MB) T-HD-04:22000(458.75 MB) T-HD-08:22000(577.37 MB)
          - FinalizationTimer: 723.747ms
         Averaged Fragment 0:(Active: 11s982ms, % non-child: 0.00%)
           split sizes: min: 384.00 MB, max: 512.00 MB, avg: 441.02 MB, stddev: 37.87 MB
           completion times: min:9s128ms max:21s068ms mean: 12s344ms stddev:2s453ms
           execution rates: min:21.26 MB/sec max:48.75 MB/sec mean:36.80 MB/sec stddev:6.37 MB/sec
           num instances: 21
            - AverageThreadTokens: 5.59
            - PeakMemoryUsage: 535.36 MB
            - RowsProduced: 563.79K (563786)
           HdfsTableSink:(Active: 11s845ms, % non-child: 100.00%)
              - EncodeTimer: 3s602ms
              - HdfsWriteTimer: 6s424ms
              - MemoryUsed: 0.00
              - RowsInserted: 563.79K (563786)
           HDFS_SCAN_NODE (id=0):(Active: 494.363ms, % non-child: 0.32%)
              - AverageHdfsReadThreadConcurrency: 0.13
              - AverageScannerThreadConcurrency: 4.72
              - BytesRead: 441.03 MB
              - BytesReadLocal: 441.02 MB
              - BytesReadShortCircuit: 441.02 MB
              - MemoryUsed: 43.76 KB
              - NumDisksAccessed: 5
              - NumScannerThreadsStarted: 7
              - PerReadThreadRawHdfsThroughput: 736.52 MB/sec
              - RowsRead: 929.16K (929159)
              - RowsReturned: 563.79K (563786)
              - RowsReturnedRate: 18.56 M/sec
              - ScanRangesComplete: 7
              - ScannerThreadsInvoluntaryContextSwitches: 12
              - ScannerThreadsTotalWallClockTime: 55s944ms
                - DelimiterParseTime: 1s395ms
                - MaterializeTupleTime(*): 1s892ms
                - ScannerThreadsSysTime: 6.520ms
                - ScannerThreadsUserTime: 3s294ms
              - ScannerThreadsVoluntaryContextSwitches: 480
              - TotalRawHdfsReadTime(*): 1s034ms
              - TotalReadThroughput: 37.91 MB/sec
         Fragment 0:
           Instance c7406bd47560e48d:e6576df75f99efac (host=T-HD-04:22000):(Active: 21s028ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/320.00 MB 1:2/128.00 MB
              - AverageThreadTokens: 4.69
              - PeakMemoryUsage: 458.75 MB
              - RowsProduced: 541.88K (541877)
             HdfsTableSink:(Active: 11s279ms, % non-child: 53.64%)
                - EncodeTimer: 3s167ms
                - HdfsWriteTimer: 6s607ms
                - MemoryUsed: 0.00
                - RowsInserted: 541.88K (541877)
             HDFS_SCAN_NODE (id=0):(Active: 9s785ms, % non-child: 46.53%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/320.00 MB 1:2/128.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:97.62% 1:0% 2:2.381%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.05
                - AverageScannerThreadConcurrency: 3.69
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 42.97 KB
                - NumDisksAccessed: 2
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 720.11 MB/sec
                - RowsRead: 942.50K (942505)
                - RowsReturned: 541.88K (541877)
                - RowsReturnedRate: 55.38 K/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 12
                - ScannerThreadsTotalWallClockTime: 1m18s
                  - DelimiterParseTime: 882.378ms
                  - MaterializeTupleTime(*): 1s519ms
                  - ScannerThreadsSysTime: 10.995ms
                  - ScannerThreadsUserTime: 2s406ms
                - ScannerThreadsVoluntaryContextSwitches: 532
                - TotalRawHdfsReadTime(*): 622.139ms
                - TotalReadThroughput: 21.33 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9b (host=T-HD-23:22000):(Active: 13s766ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:2/128.00 MB 9:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 7.19
              - PeakMemoryUsage: 592.74 MB
              - RowsProduced: 742.68K (742677)
             HdfsTableSink:(Active: 13s773ms, % non-child: 100.00%)
                - EncodeTimer: 4s514ms
                - HdfsWriteTimer: 7s758ms
                - MemoryUsed: 0.00
                - RowsInserted: 742.68K (742677)
             HDFS_SCAN_NODE (id=0):(Active: 28.677ms, % non-child: 0.21%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:2/128.00 MB 9:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 6.19
                - BytesRead: 512.01 MB
                - BytesReadLocal: 512.00 MB
                - BytesReadShortCircuit: 512.00 MB
                - MemoryUsed: 51.08 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 945.69 MB/sec
                - RowsRead: 1.08M (1082397)
                - RowsReturned: 742.68K (742677)
                - RowsReturnedRate: 25.90 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 19
                - ScannerThreadsTotalWallClockTime: 1m25s
                  - DelimiterParseTime: 1s941ms
                  - MaterializeTupleTime(*): 2s380ms
                  - ScannerThreadsSysTime: 6.995ms
                  - ScannerThreadsUserTime: 4s336ms
                - ScannerThreadsVoluntaryContextSwitches: 653
                - TotalRawHdfsReadTime(*): 541.413ms
                - TotalReadThroughput: 37.90 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa1 (host=T-HD-11:22000):(Active: 13s649ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 2:1/64.00 MB 3:1/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 4.44
              - PeakMemoryUsage: 472.98 MB
              - RowsProduced: 520.73K (520730)
             HdfsTableSink:(Active: 13s648ms, % non-child: 99.99%)
                - EncodeTimer: 3s391ms
                - HdfsWriteTimer: 8s770ms
                - MemoryUsed: 0.00
                - RowsInserted: 520.73K (520730)
             HDFS_SCAN_NODE (id=0):(Active: 34.308ms, % non-child: 0.25%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 2:1/64.00 MB 3:1/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:96.3% 1:3.704% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:6
               ExecOption: Codegen enabled: 0 out of 6
                - AverageHdfsReadThreadConcurrency: 0.04
                - AverageScannerThreadConcurrency: 3.44
                - BytesRead: 384.01 MB
                - BytesReadLocal: 384.00 MB
                - BytesReadShortCircuit: 384.00 MB
                - MemoryUsed: 36.95 KB
                - NumDisksAccessed: 5
                - NumScannerThreadsStarted: 6
                - PerReadThreadRawHdfsThroughput: 965.65 MB/sec
                - RowsRead: 791.98K (791984)
                - RowsReturned: 520.73K (520730)
                - RowsReturnedRate: 15.18 M/sec
                - ScanRangesComplete: 6
                - ScannerThreadsInvoluntaryContextSwitches: 20
                - ScannerThreadsTotalWallClockTime: 47s017ms
                  - DelimiterParseTime: 1s206ms
                  - MaterializeTupleTime(*): 1s677ms
                  - ScannerThreadsSysTime: 8.997ms
                  - ScannerThreadsUserTime: 2s885ms
                - ScannerThreadsVoluntaryContextSwitches: 418
                - TotalRawHdfsReadTime(*): 397.666ms
                - TotalReadThroughput: 28.43 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa2 (host=T-HD-20:22000):(Active: 13s416ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:2/128.00 MB 5:1/64.00 MB 6:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 6.59
              - PeakMemoryUsage: 520.36 MB
              - RowsProduced: 621.47K (621467)
             HdfsTableSink:(Active: 13s430ms, % non-child: 100.00%)
                - EncodeTimer: 3s742ms
                - HdfsWriteTimer: 8s181ms
                - MemoryUsed: 0.00
                - RowsInserted: 621.47K (621467)
             HDFS_SCAN_NODE (id=0):(Active: 28.330ms, % non-child: 0.21%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:2/128.00 MB 5:1/64.00 MB 6:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 5.59
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 44.25 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 854.55 MB/sec
                - RowsRead: 959.60K (959598)
                - RowsReturned: 621.47K (621467)
                - RowsReturnedRate: 21.94 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 9
                - ScannerThreadsTotalWallClockTime: 1m15s
                  - DelimiterParseTime: 1s610ms
                  - MaterializeTupleTime(*): 2s062ms
                  - ScannerThreadsSysTime: 3.997ms
                  - ScannerThreadsUserTime: 3s689ms
                - ScannerThreadsVoluntaryContextSwitches: 522
                - TotalRawHdfsReadTime(*): 524.262ms
                - TotalReadThroughput: 33.18 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa6 (host=T-HD-18:22000):(Active: 13s277ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 3:1/64.00 MB 6:2/113.42 MB 7:2/128.00 MB 9:1/62.04 MB
              - AverageThreadTokens: 5.00
              - PeakMemoryUsage: 541.48 MB
              - RowsProduced: 634.46K (634456)
             HdfsTableSink:(Active: 13s370ms, % non-child: 100.00%)
                - EncodeTimer: 4s177ms
                - HdfsWriteTimer: 7s603ms
                - MemoryUsed: 0.00
                - RowsInserted: 634.46K (634456)
             HDFS_SCAN_NODE (id=0):(Active: 37.367ms, % non-child: 0.28%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 3:1/64.00 MB 6:2/113.42 MB 7:2/128.00 MB 9:1/62.04 MB
               Hdfs Read Thread Concurrency Bucket: 0:85.19% 1:3.704% 2:0% 3:0% 4:11.11% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.48
                - AverageScannerThreadConcurrency: 4.00
                - BytesRead: 431.47 MB
                - BytesReadLocal: 431.47 MB
                - BytesReadShortCircuit: 431.47 MB
                - MemoryUsed: 39.16 KB
                - NumDisksAccessed: 5
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 72.12 MB/sec
                - RowsRead: 908.73K (908733)
                - RowsReturned: 634.46K (634456)
                - RowsReturnedRate: 16.98 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 11
                - ScannerThreadsTotalWallClockTime: 53s134ms
                  - DelimiterParseTime: 1s464ms
                  - MaterializeTupleTime(*): 1s938ms
                  - ScannerThreadsSysTime: 2.998ms
                  - ScannerThreadsUserTime: 3s401ms
                - ScannerThreadsVoluntaryContextSwitches: 518
                - TotalRawHdfsReadTime(*): 5s982ms
                - TotalReadThroughput: 31.93 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9d (host=T-HD-06:22000):(Active: 12s990ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 2:1/64.00 MB 3:1/64.00 MB 5:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB
              - AverageThreadTokens: 5.41
              - PeakMemoryUsage: 552.36 MB
              - RowsProduced: 601.57K (601570)
             HdfsTableSink:(Active: 13s396ms, % non-child: 100.00%)
                - EncodeTimer: 3s915ms
                - HdfsWriteTimer: 7s597ms
                - MemoryUsed: 0.00
                - RowsInserted: 601.57K (601570)
             HDFS_SCAN_NODE (id=0):(Active: 30.702ms, % non-child: 0.24%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 2:1/64.00 MB 3:1/64.00 MB 5:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:96.15% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:3.846% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.27
                - AverageScannerThreadConcurrency: 4.58
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 46.37 KB
                - NumDisksAccessed: 7
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 842.59 MB/sec
                - RowsRead: 931.10K (931097)
                - RowsReturned: 601.57K (601570)
                - RowsReturnedRate: 19.59 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 18
                - ScannerThreadsTotalWallClockTime: 58s127ms
                  - DelimiterParseTime: 1s589ms
                  - MaterializeTupleTime(*): 2s038ms
                  - ScannerThreadsSysTime: 3.997ms
                  - ScannerThreadsUserTime: 3s629ms
                - ScannerThreadsVoluntaryContextSwitches: 517
                - TotalRawHdfsReadTime(*): 531.705ms
                - TotalReadThroughput: 34.42 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efad (host=T-HD-07:22000):(Active: 12s767ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 3:1/64.00 MB 5:3/192.00 MB 6:1/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 6.81
              - PeakMemoryUsage: 616.75 MB
              - RowsProduced: 668.95K (668951)
             HdfsTableSink:(Active: 13s326ms, % non-child: 100.00%)
                - EncodeTimer: 3s996ms
                - HdfsWriteTimer: 7s276ms
                - MemoryUsed: 0.00
                - RowsInserted: 668.95K (668951)
             HDFS_SCAN_NODE (id=0):(Active: 26.716ms, % non-child: 0.21%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 3:1/64.00 MB 5:3/192.00 MB 6:1/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:84.62% 1:3.846% 2:7.692% 3:0% 4:3.846% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.35
                - AverageScannerThreadConcurrency: 6.04
                - BytesRead: 512.01 MB
                - BytesReadLocal: 512.00 MB
                - BytesReadShortCircuit: 512.00 MB
                - MemoryUsed: 50.46 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 154.41 MB/sec
                - RowsRead: 1.07M (1065500)
                - RowsReturned: 668.95K (668951)
                - RowsReturnedRate: 25.04 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 30
                - ScannerThreadsTotalWallClockTime: 1m17s
                  - DelimiterParseTime: 1s737ms
                  - MaterializeTupleTime(*): 2s326ms
                  - ScannerThreadsSysTime: 9.994ms
                  - ScannerThreadsUserTime: 4s058ms
                - ScannerThreadsVoluntaryContextSwitches: 570
                - TotalRawHdfsReadTime(*): 3s315ms
                - TotalReadThroughput: 39.33 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa5 (host=T-HD-03:22000):(Active: 12s716ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/261.66 MB 1:2/128.00 MB
              - AverageThreadTokens: 5.96
              - PeakMemoryUsage: 410.25 MB
              - RowsProduced: 472.73K (472725)
             HdfsTableSink:(Active: 13s339ms, % non-child: 100.00%)
                - EncodeTimer: 2s777ms
                - HdfsWriteTimer: 8s436ms
                - MemoryUsed: 0.00
                - RowsInserted: 472.73K (472725)
             HDFS_SCAN_NODE (id=0):(Active: 23.337ms, % non-child: 0.18%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/261.66 MB 1:2/128.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:84.62% 1:11.54% 2:3.846%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.19
                - AverageScannerThreadConcurrency: 5.15
                - BytesRead: 389.67 MB
                - BytesReadLocal: 389.66 MB
                - BytesReadShortCircuit: 389.66 MB
                - MemoryUsed: 39.62 KB
                - NumDisksAccessed: 2
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 229.99 MB/sec
                - RowsRead: 812.09K (812091)
                - RowsReturned: 472.73K (472725)
                - RowsReturnedRate: 20.26 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 5
                - ScannerThreadsTotalWallClockTime: 1m5s
                  - DelimiterParseTime: 773.201ms
                  - MaterializeTupleTime(*): 1s320ms
                  - ScannerThreadsSysTime: 2.997ms
                  - ScannerThreadsUserTime: 2s092ms
                - ScannerThreadsVoluntaryContextSwitches: 468
                - TotalRawHdfsReadTime(*): 1s694ms
                - TotalReadThroughput: 29.97 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa4 (host=T-HD-16:22000):(Active: 12s146ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 2:1/64.00 MB 3:1/64.00 MB 6:1/64.00 MB 7:2/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 4.29
              - PeakMemoryUsage: 569.36 MB
              - RowsProduced: 585.87K (585865)
             HdfsTableSink:(Active: 12s424ms, % non-child: 100.00%)
                - EncodeTimer: 3s961ms
                - HdfsWriteTimer: 6s694ms
                - MemoryUsed: 0.00
                - RowsInserted: 585.87K (585865)
             HDFS_SCAN_NODE (id=0):(Active: 28.949ms, % non-child: 0.24%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 2:1/64.00 MB 3:1/64.00 MB 6:1/64.00 MB 7:2/64.00 MB 8:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 3.29
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 44.53 KB
                - NumDisksAccessed: 7
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 945.31 MB/sec
                - RowsRead: 927.24K (927240)
                - RowsReturned: 585.87K (585865)
                - RowsReturnedRate: 20.24 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 10
                - ScannerThreadsTotalWallClockTime: 40s968ms
                  - DelimiterParseTime: 1s521ms
                  - MaterializeTupleTime(*): 1s961ms
                  - ScannerThreadsSysTime: 2.997ms
                  - ScannerThreadsUserTime: 3s494ms
                - ScannerThreadsVoluntaryContextSwitches: 504
                - TotalRawHdfsReadTime(*): 473.925ms
                - TotalReadThroughput: 37.32 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efaa (host=T-HD-14:22000):(Active: 11s816ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 3:1/64.00 MB 4:2/126.93 MB 5:1/64.00 MB 7:2/128.00 MB 9:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 6.08
              - PeakMemoryUsage: 642.26 MB
              - RowsProduced: 589.88K (589878)
             HdfsTableSink:(Active: 12s411ms, % non-child: 100.00%)
                - EncodeTimer: 3s901ms
                - HdfsWriteTimer: 6s417ms
                - MemoryUsed: 0.00
                - RowsInserted: 589.88K (589878)
             HDFS_SCAN_NODE (id=0):(Active: 33.68ms, % non-child: 0.28%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 3:1/64.00 MB 4:2/126.93 MB 5:1/64.00 MB 7:2/128.00 MB 9:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:95.83% 1:4.167% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.04
                - AverageScannerThreadConcurrency: 5.29
                - BytesRead: 510.93 MB
                - BytesReadLocal: 510.93 MB
                - BytesReadShortCircuit: 510.93 MB
                - MemoryUsed: 51.30 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 909.30 MB/sec
                - RowsRead: 1.07M (1067212)
                - RowsReturned: 589.88K (589878)
                - RowsReturnedRate: 17.84 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 10
                - ScannerThreadsTotalWallClockTime: 1m2s
                  - DelimiterParseTime: 1s560ms
                  - MaterializeTupleTime(*): 2s205ms
                  - ScannerThreadsSysTime: 5.996ms
                  - ScannerThreadsUserTime: 3s789ms
                - ScannerThreadsVoluntaryContextSwitches: 479
                - TotalRawHdfsReadTime(*): 561.900ms
                - TotalReadThroughput: 42.56 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efab (host=T-HD-13:22000):(Active: 11s758ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 2:2/128.00 MB 4:2/128.00 MB 6:2/128.00 MB 7:1/64.00 MB
              - AverageThreadTokens: 4.68
              - PeakMemoryUsage: 582.48 MB
              - RowsProduced: 556.84K (556842)
             HdfsTableSink:(Active: 12s415ms, % non-child: 100.00%)
                - EncodeTimer: 3s670ms
                - HdfsWriteTimer: 6s591ms
                - MemoryUsed: 0.00
                - RowsInserted: 556.84K (556842)
             HDFS_SCAN_NODE (id=0):(Active: 29.52ms, % non-child: 0.25%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 2:2/128.00 MB 4:2/128.00 MB 6:2/128.00 MB 7:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 4.00
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 38.04 KB
                - NumDisksAccessed: 4
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 977.48 MB/sec
                - RowsRead: 966.68K (966683)
                - RowsReturned: 556.84K (556842)
                - RowsReturnedRate: 19.17 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 6
                - ScannerThreadsTotalWallClockTime: 47s013ms
                  - DelimiterParseTime: 1s448ms
                  - MaterializeTupleTime(*): 1s982ms
                  - ScannerThreadsSysTime: 999.0us
                  - ScannerThreadsUserTime: 3s443ms
                - ScannerThreadsVoluntaryContextSwitches: 452
                - TotalRawHdfsReadTime(*): 458.328ms
                - TotalReadThroughput: 38.93 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa7 (host=T-HD-17:22000):(Active: 11s655ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 3:1/64.00 MB 4:2/128.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB
              - AverageThreadTokens: 5.44
              - PeakMemoryUsage: 552.86 MB
              - RowsProduced: 558.02K (558025)
             HdfsTableSink:(Active: 12s421ms, % non-child: 100.00%)
                - EncodeTimer: 3s636ms
                - HdfsWriteTimer: 6s520ms
                - MemoryUsed: 0.00
                - RowsInserted: 558.02K (558025)
             HDFS_SCAN_NODE (id=0):(Active: 26.444ms, % non-child: 0.23%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 3:1/64.00 MB 4:2/128.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 4.83
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 45.52 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 894.03 MB/sec
                - RowsRead: 963.92K (963924)
                - RowsReturned: 558.02K (558025)
                - RowsReturnedRate: 21.10 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 15
                - ScannerThreadsTotalWallClockTime: 56s292ms
                  - DelimiterParseTime: 1s526ms
                  - MaterializeTupleTime(*): 1s952ms
                  - ScannerThreadsSysTime: 999.0us
                  - ScannerThreadsUserTime: 3s496ms
                - ScannerThreadsVoluntaryContextSwitches: 458
                - TotalRawHdfsReadTime(*): 501.110ms
                - TotalReadThroughput: 38.94 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa9 (host=T-HD-21:22000):(Active: 11s043ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 4:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 3.92
              - PeakMemoryUsage: 504.48 MB
              - RowsProduced: 487.26K (487263)
             HdfsTableSink:(Active: 11s895ms, % non-child: 100.00%)
                - EncodeTimer: 3s176ms
                - HdfsWriteTimer: 6s353ms
                - MemoryUsed: 0.00
                - RowsInserted: 487.26K (487263)
             HDFS_SCAN_NODE (id=0):(Active: 26.675ms, % non-child: 0.24%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 4:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:90.91% 1:4.545% 2:0% 3:0% 4:4.545% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:6
               ExecOption: Codegen enabled: 0 out of 6
                - AverageHdfsReadThreadConcurrency: 0.23
                - AverageScannerThreadConcurrency: 3.18
                - BytesRead: 384.01 MB
                - BytesReadLocal: 384.00 MB
                - BytesReadShortCircuit: 384.00 MB
                - MemoryUsed: 36.85 KB
                - NumDisksAccessed: 5
                - NumScannerThreadsStarted: 6
                - PerReadThreadRawHdfsThroughput: 922.42 MB/sec
                - RowsRead: 806.12K (806120)
                - RowsReturned: 487.26K (487263)
                - RowsReturnedRate: 18.27 M/sec
                - ScanRangesComplete: 6
                - ScannerThreadsInvoluntaryContextSwitches: 6
                - ScannerThreadsTotalWallClockTime: 33s986ms
                  - DelimiterParseTime: 1s209ms
                  - MaterializeTupleTime(*): 1s650ms
                  - ScannerThreadsSysTime: 8.995ms
                  - ScannerThreadsUserTime: 2s855ms
                - ScannerThreadsVoluntaryContextSwitches: 387
                - TotalRawHdfsReadTime(*): 416.301ms
                - TotalReadThroughput: 34.90 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa0 (host=T-HD-15:22000):(Active: 11s000ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 1:1/64.00 MB 4:1/64.00 MB 5:1/64.00 MB 6:1/64.00 MB 7:2/64.36 MB 9:1/64.00 MB
              - AverageThreadTokens: 5.48
              - PeakMemoryUsage: 544.86 MB
              - RowsProduced: 595.95K (595950)
             HdfsTableSink:(Active: 11s608ms, % non-child: 100.00%)
                - EncodeTimer: 3s974ms
                - HdfsWriteTimer: 5s550ms
                - MemoryUsed: 0.00
                - RowsInserted: 595.95K (595950)
             HDFS_SCAN_NODE (id=0):(Active: 24.960ms, % non-child: 0.23%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 1:1/64.00 MB 4:1/64.00 MB 5:1/64.00 MB 6:1/64.00 MB 7:2/64.36 MB 9:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 4.68
                - BytesRead: 448.37 MB
                - BytesReadLocal: 448.36 MB
                - BytesReadShortCircuit: 448.36 MB
                - MemoryUsed: 46.95 KB
                - NumDisksAccessed: 7
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 842.57 MB/sec
                - RowsRead: 933.58K (933577)
                - RowsReturned: 595.95K (595950)
                - RowsReturnedRate: 23.88 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 21
                - ScannerThreadsTotalWallClockTime: 51s278ms
                  - DelimiterParseTime: 1s599ms
                  - MaterializeTupleTime(*): 2s052ms
                  - ScannerThreadsSysTime: 17.994ms
                  - ScannerThreadsUserTime: 3s648ms
                - ScannerThreadsVoluntaryContextSwitches: 532
                - TotalRawHdfsReadTime(*): 532.137ms
                - TotalReadThroughput: 40.75 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9e (host=T-HD-09:22000):(Active: 10s651ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 2:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB 10:2/128.00 MB
              - AverageThreadTokens: 5.82
              - PeakMemoryUsage: 551.36 MB
              - RowsProduced: 526.81K (526806)
             HdfsTableSink:(Active: 11s022ms, % non-child: 100.00%)
                - EncodeTimer: 3s402ms
                - HdfsWriteTimer: 5s761ms
                - MemoryUsed: 0.00
                - RowsInserted: 526.81K (526806)
             HDFS_SCAN_NODE (id=0):(Active: 26.834ms, % non-child: 0.25%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 2:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB 10:2/128.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 5.05
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 45.38 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 813.45 MB/sec
                - RowsRead: 937.30K (937302)
                - RowsReturned: 526.81K (526806)
                - RowsReturnedRate: 19.63 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 13
                - ScannerThreadsTotalWallClockTime: 53s054ms
                  - DelimiterParseTime: 1s445ms
                  - MaterializeTupleTime(*): 1s938ms
                  - ScannerThreadsSysTime: 7.995ms
                  - ScannerThreadsUserTime: 3s391ms
                - ScannerThreadsVoluntaryContextSwitches: 422
                - TotalRawHdfsReadTime(*): 550.750ms
                - TotalReadThroughput: 42.64 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa8 (host=T-HD-02:22000):(Active: 10s429ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/267.28 MB 1:2/128.00 MB
              - AverageThreadTokens: 6.52
              - PeakMemoryUsage: 394.25 MB
              - RowsProduced: 540.27K (540271)
             HdfsTableSink:(Active: 10s530ms, % non-child: 100.00%)
                - EncodeTimer: 3s163ms
                - HdfsWriteTimer: 5s759ms
                - MemoryUsed: 0.00
                - RowsInserted: 540.27K (540271)
             HDFS_SCAN_NODE (id=0):(Active: 54.235ms, % non-child: 0.52%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:5/267.28 MB 1:2/128.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:95.24% 1:4.762% 2:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.05
                - AverageScannerThreadConcurrency: 5.52
                - BytesRead: 395.28 MB
                - BytesReadLocal: 395.28 MB
                - BytesReadShortCircuit: 395.28 MB
                - MemoryUsed: 39.20 KB
                - NumDisksAccessed: 2
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 709.27 MB/sec
                - RowsRead: 821.09K (821089)
                - RowsReturned: 540.27K (540271)
                - RowsReturnedRate: 9.96 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 17
                - ScannerThreadsTotalWallClockTime: 57s619ms
                  - DelimiterParseTime: 782.499ms
                  - MaterializeTupleTime(*): 1s339ms
                  - ScannerThreadsSysTime: 8.996ms
                  - ScannerThreadsUserTime: 2s130ms
                - ScannerThreadsVoluntaryContextSwitches: 533
                - TotalRawHdfsReadTime(*): 557.307ms
                - TotalReadThroughput: 37.63 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9f (host=T-HD-24:22000):(Active: 10s253ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/111.47 MB 1:1/64.00 MB 6:1/64.00 MB 7:2/65.18 MB 9:1/64.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 6.38
              - PeakMemoryUsage: 553.86 MB
              - RowsProduced: 448.16K (448158)
             HdfsTableSink:(Active: 10s348ms, % non-child: 100.00%)
                - EncodeTimer: 2s935ms
                - HdfsWriteTimer: 5s840ms
                - MemoryUsed: 0.00
                - RowsInserted: 448.16K (448158)
             HDFS_SCAN_NODE (id=0):(Active: 25.29ms, % non-child: 0.24%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/111.47 MB 1:1/64.00 MB 6:1/64.00 MB 7:2/65.18 MB 9:1/64.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:8
               ExecOption: Codegen enabled: 0 out of 8
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 5.38
                - BytesRead: 432.66 MB
                - BytesReadLocal: 432.66 MB
                - BytesReadShortCircuit: 432.66 MB
                - MemoryUsed: 47.30 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 8
                - PerReadThreadRawHdfsThroughput: 937.53 MB/sec
                - RowsRead: 916.98K (916978)
                - RowsReturned: 448.16K (448158)
                - RowsReturnedRate: 17.91 M/sec
                - ScanRangesComplete: 8
                - ScannerThreadsInvoluntaryContextSwitches: 9
                - ScannerThreadsTotalWallClockTime: 56s498ms
                  - DelimiterParseTime: 1s410ms
                  - MaterializeTupleTime(*): 1s756ms
                  - ScannerThreadsSysTime: 5.995ms
                  - ScannerThreadsUserTime: 3s179ms
                - ScannerThreadsVoluntaryContextSwitches: 351
                - TotalRawHdfsReadTime(*): 461.489ms
                - TotalReadThroughput: 41.20 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efa3 (host=T-HD-19:22000):(Active: 9s923ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 2:1/64.00 MB 3:1/64.00 MB 5:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB
              - AverageThreadTokens: 6.00
              - PeakMemoryUsage: 480.98 MB
              - RowsProduced: 496.62K (496625)
             HdfsTableSink:(Active: 9s924ms, % non-child: 100.00%)
                - EncodeTimer: 3s295ms
                - HdfsWriteTimer: 5s128ms
                - MemoryUsed: 0.00
                - RowsInserted: 496.62K (496625)
             HDFS_SCAN_NODE (id=0):(Active: 27.832ms, % non-child: 0.28%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:1/64.00 MB 2:1/64.00 MB 3:1/64.00 MB 5:1/64.00 MB 8:1/64.00 MB 9:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:95% 1:0% 2:0% 3:0% 4:0% 5:0% 6:5% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:6
               ExecOption: Codegen enabled: 0 out of 6
                - AverageHdfsReadThreadConcurrency: 0.30
                - AverageScannerThreadConcurrency: 5.00
                - BytesRead: 384.01 MB
                - BytesReadLocal: 384.00 MB
                - BytesReadShortCircuit: 384.00 MB
                - MemoryUsed: 35.94 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 6
                - PerReadThreadRawHdfsThroughput: 847.29 MB/sec
                - RowsRead: 831.10K (831100)
                - RowsReturned: 496.62K (496625)
                - RowsReturnedRate: 17.84 M/sec
                - ScanRangesComplete: 6
                - ScannerThreadsInvoluntaryContextSwitches: 7
                - ScannerThreadsTotalWallClockTime: 47s761ms
                  - DelimiterParseTime: 1s193ms
                  - MaterializeTupleTime(*): 1s703ms
                  - ScannerThreadsSysTime: 6.996ms
                  - ScannerThreadsUserTime: 2s904ms
                - ScannerThreadsVoluntaryContextSwitches: 397
                - TotalRawHdfsReadTime(*): 453.217ms
                - TotalReadThroughput: 38.35 MB/sec
           Instance c7406bd47560e48d:e6576df75f99efae (host=T-HD-08:22000):(Active: 9s471ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 4:1/64.00 MB 7:2/128.00 MB 8:2/128.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 5.00
              - PeakMemoryUsage: 577.37 MB
              - RowsProduced: 554.99K (554993)
             HdfsTableSink:(Active: 9s483ms, % non-child: 100.00%)
                - EncodeTimer: 3s666ms
                - HdfsWriteTimer: 4s307ms
                - MemoryUsed: 0.00
                - RowsInserted: 554.99K (554993)
             HDFS_SCAN_NODE (id=0):(Active: 28.586ms, % non-child: 0.30%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 1:1/64.00 MB 4:1/64.00 MB 7:2/128.00 MB 8:2/128.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:84.21% 1:10.53% 2:0% 3:0% 4:0% 5:5.263% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.37
                - AverageScannerThreadConcurrency: 4.00
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 41.49 KB
                - NumDisksAccessed: 5
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 215.07 MB/sec
                - RowsRead: 966.71K (966705)
                - RowsReturned: 554.99K (554993)
                - RowsReturnedRate: 19.41 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 4
                - ScannerThreadsTotalWallClockTime: 36s823ms
                  - DelimiterParseTime: 1s392ms
                  - MaterializeTupleTime(*): 1s899ms
                  - ScannerThreadsSysTime: 3.998ms
                  - ScannerThreadsUserTime: 3s311ms
                - ScannerThreadsVoluntaryContextSwitches: 500
                - TotalRawHdfsReadTime(*): 2s083ms
                - TotalReadThroughput: 47.13 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9a (host=T-HD-10:22000):(Active: 9s075ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 2:1/64.00 MB 4:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB
              - AverageThreadTokens: 5.79
              - PeakMemoryUsage: 561.36 MB
              - RowsProduced: 537.36K (537363)
             HdfsTableSink:(Active: 9s599ms, % non-child: 100.00%)
                - EncodeTimer: 3s531ms
                - HdfsWriteTimer: 4s054ms
                - MemoryUsed: 0.00
                - RowsInserted: 537.36K (537363)
             HDFS_SCAN_NODE (id=0):(Active: 28.676ms, % non-child: 0.32%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 0:2/128.00 MB 2:1/64.00 MB 4:1/64.00 MB 6:1/64.00 MB 7:1/64.00 MB 8:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:100% 1:0% 2:0% 3:0% 4:0% 5:0% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.00
                - AverageScannerThreadConcurrency: 5.06
                - BytesRead: 448.01 MB
                - BytesReadLocal: 448.00 MB
                - BytesReadShortCircuit: 448.00 MB
                - MemoryUsed: 46.84 KB
                - NumDisksAccessed: 6
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 808.87 MB/sec
                - RowsRead: 942.39K (942387)
                - RowsReturned: 537.36K (537363)
                - RowsReturnedRate: 18.74 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 9
                - ScannerThreadsTotalWallClockTime: 46s899ms
                  - DelimiterParseTime: 1s596ms
                  - MaterializeTupleTime(*): 1s987ms
                  - ScannerThreadsSysTime: 5.996ms
                  - ScannerThreadsUserTime: 3s588ms
                - ScannerThreadsVoluntaryContextSwitches: 426
                - TotalRawHdfsReadTime(*): 553.865ms
                - TotalReadThroughput: 49.76 MB/sec
           Instance c7406bd47560e48d:e6576df75f99ef9c (host=T-HD-22:22000):(Active: 8s795ms, % non-child: 0.00%)
             Hdfs split stats (<volume id>:<# splits>/<split lengths>): 4:1/64.00 MB 5:1/64.00 MB 8:2/125.07 MB 9:2/128.00 MB 10:1/64.00 MB
              - AverageThreadTokens: 5.79
              - PeakMemoryUsage: 561.36 MB
              - RowsProduced: 557.02K (557024)
             HdfsTableSink:(Active: 9s097ms, % non-child: 100.00%)
                - EncodeTimer: 3s649ms
                - HdfsWriteTimer: 3s697ms
                - MemoryUsed: 0.00
                - RowsInserted: 557.02K (557024)
             HDFS_SCAN_NODE (id=0):(Active: 26.763ms, % non-child: 0.30%)
               Hdfs split stats (<volume id>:<# splits>/<split lengths>): 4:1/64.00 MB 5:1/64.00 MB 8:2/125.07 MB 9:2/128.00 MB 10:1/64.00 MB
               Hdfs Read Thread Concurrency Bucket: 0:88.89% 1:5.556% 2:0% 3:0% 4:0% 5:5.556% 6:0% 7:0% 8:0% 9:0% 10:0% 11:0% 12:0% 13:0%
               File Formats: TEXT/NONE:7
               ExecOption: Codegen enabled: 0 out of 7
                - AverageHdfsReadThreadConcurrency: 0.33
                - AverageScannerThreadConcurrency: 5.06
                - BytesRead: 445.07 MB
                - BytesReadLocal: 445.07 MB
                - BytesReadShortCircuit: 445.07 MB
                - MemoryUsed: 48.74 KB
                - NumDisksAccessed: 5
                - NumScannerThreadsStarted: 7
                - PerReadThreadRawHdfsThroughput: 859.30 MB/sec
                - RowsRead: 938.13K (938135)
                - RowsReturned: 557.02K (557024)
                - RowsReturnedRate: 20.81 M/sec
                - ScanRangesComplete: 7
                - ScannerThreadsInvoluntaryContextSwitches: 12
                - ScannerThreadsTotalWallClockTime: 43s410ms
                  - DelimiterParseTime: 1s408ms
                  - MaterializeTupleTime(*): 2s047ms
                  - ScannerThreadsSysTime: 7.994ms
                  - ScannerThreadsUserTime: 3s459ms
                - ScannerThreadsVoluntaryContextSwitches: 451
                - TotalRawHdfsReadTime(*): 517.951ms
                - TotalReadThroughput: 49.41 MB/sec

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupimpala-user @
categorieshadoop
postedAug 9, '13 at 2:36a
activeAug 12, '13 at 7:39a
posts10
users3
websitecloudera.com
irc#hadoop

3 users in discussion

邱睿: 6 posts Greg Rahn: 2 posts Alex Behm: 2 posts

People

Translate

site design / logo © 2021 Grokbase