Search Discussions

30 discussions - 90 posts

  • We're running some benchmarks on Impala 2.0. In the first part of the blog, we have a detailed look at how Cloudera ran their previous benchmarks ...
    Kris PeetersKris Peeters
    Nov 2, 2014 at 11:42 am
    Nov 4, 2014 at 9:45 am
  • Hello I have a question about different partitions listed in the query plan. I'm not sure my understanding is correct, I appreciate if anyone understands the query plane well, can help me figure ...
    Nov 10, 2014 at 9:39 am
    Nov 10, 2014 at 11:01 am
  • Hi Dima - When an Impala daemon goes offline, the following should happen: 1. The statestore should detect that its regular heartbeats are not getting through to the daemon. When that happens, the ...
    Henry RobinsonHenry Robinson
    Nov 3, 2014 at 5:46 pm
    Nov 12, 2014 at 7:28 pm
  • Ken, the date_id approach requires dynamic partition pruning in order to avoid having to scan the entire table for each query. Impala doesn't support that at the moment, but this is on our ...
    Marcel KornackerMarcel Kornacker
    Nov 6, 2014 at 10:20 pm
    Nov 12, 2014 at 11:24 am
  • Hello, I'm trying to enable resource management on Impala using Llama on Yarn. (Running CDH5.2 and CM) I used the "wizard" from the pool's webpage to do so. After a restart of the services, ...
    Philippe MarseillePhilippe Marseille
    Nov 7, 2014 at 8:08 pm
    Nov 11, 2014 at 9:21 pm
  • Hi, I tried play with UDAF collecting numeric from all rows in array and doing some calculation in Finalise step (probably not the best approach in distributed environment but should be feasible for ...
    Jonas JarutisJonas Jarutis
    Nov 6, 2014 at 8:26 am
    Nov 10, 2014 at 7:55 am
  • Stephane, I would take advantage of the fact that you can point external impala tables to new locations while online (in fact multiple tables to same location). So you could keep copy A under ...
    Mauricio AristizabalMauricio Aristizabal
    Nov 4, 2014 at 6:30 pm
    Nov 5, 2014 at 4:20 pm
  • Hi all, I have created a simple UDF AddUdf on my redhat box. Instead of IR, I created a shared object of this UDF and transferred it to CDH 5.2 quick VM with Impala 2.0. While I am trying to create ...
    Shabnam perweenShabnam perween
    Nov 11, 2014 at 3:58 am
    Nov 12, 2014 at 5:28 am
  • Hi there, I have been looking into Impala 2.0.0 source code recently. I saw that Impala is using libhdfs to read/write HDFS. However I didn't see the creation of directories for databases/tables in ...
    Chengbing LiuChengbing Liu
    Nov 10, 2014 at 10:20 am
    Nov 12, 2014 at 1:57 am
  • Hi, In Impala 1.4, I can set parquet compression by 'set PARQUET_COMPRESSION_CODEC=snappy;' After upgraded to Impala 2.0, it reported error message as follows: Unknown query option ...
    Zhe LiZhe Li
    Nov 10, 2014 at 9:03 am
    Nov 11, 2014 at 3:28 am
  • Hi All: I have done some TPC-H tests for impala 2.0. I have found one misleading point. For example the TPC-H Query3, the total duration is 28s, but the total seconds that sums from summary is only ...
    Nov 9, 2014 at 9:02 am
    Nov 10, 2014 at 5:41 am
  • Hi Sean, Which version of Impala are you using? Can you please send the full query profile for the query that is failing as well as the impalad log? Thanks, Dimitris To unsubscribe from this group ...
    Dimitris TsirogiannisDimitris Tsirogiannis
    Nov 11, 2014 at 10:49 pm
    Nov 11, 2014 at 11:20 pm
  • I'm using CDH 5.2.0 and wanted to know if there are there is a good collection of UDFs out there that work with Impala? Is it still true that Impala doesn't support nested or composite types? Thanks! ...
    Buntu DevBuntu Dev
    Nov 11, 2014 at 6:44 pm
    Nov 11, 2014 at 6:50 pm
  • Hi People, I'm using impala 2.0.0 version. There is weird thing happening around in subqueries. I've two table name as "a" and "b": Query: select * from a +----+ +----+ Query: select * from b +----+ ...
    Ravi SharmaRavi Sharma
    Nov 11, 2014 at 10:02 am
    Nov 11, 2014 at 2:46 pm
  • Hi For a PlanFragment root, my understanding is that it is executed on the Coordinator Node, so it should only run on one host. However, the plan I got from profile indicating that it is running on 4 ...
    Nov 10, 2014 at 9:42 am
    Nov 11, 2014 at 2:58 am
  • Hi, We have an ETL process that generates parquet files every so often using spark job. I would like to use these files with an external partitioned impala table. Impala can read the original file ...
    Stephane DrouinStephane Drouin
    Nov 10, 2014 at 2:36 am
    Nov 10, 2014 at 5:44 am
  • Hi all, We'd like to give everyone a heads up on IMPALA-1401 that affects CDH 5.2. Due to this bug, Parquet files created outside Impala (eg Hive or MR) in CDH 5.2 with strings 36 bytes may fail in ...
    Justin EricksonJustin Erickson
    Nov 7, 2014 at 2:05 am
    Nov 9, 2014 at 3:10 am
  • I am pouring through the documentation for CDH5.2 and can no longer find the steps to configure Impala/Llama to use YARN along with the YARN configurations to accommodate. I see in the Static Service ...
    Benjamin KimBenjamin Kim
    Nov 7, 2014 at 5:47 pm
    Nov 7, 2014 at 6:35 pm
  • Hi, I am able to load upto 30 columns in hbase. Regex is working well. I have searched but didnt get this limit. Why am i unable to upload only upto 30 columns not more then that. Is their any limit ...
    Ashish sanadhyaAshish sanadhya
    Nov 3, 2014 at 7:18 am
    Nov 3, 2014 at 3:08 pm
  • Hi Jung-Yup, Changing the query processing model is not in the immediate plans for Impala (see http://blog.cloudera.com/blog/2014/08/whats-next-for-impala-focus-on-advanced-sql-functionality/). That ...
    Dimitris TsirogiannisDimitris Tsirogiannis
    Nov 12, 2014 at 5:48 pm
    Nov 12, 2014 at 5:48 pm
  • I saw Impala's 2.1 roadmap: - Parquet enhancements – continued performance gains including index pages Will Impala use indexes for Parquet files? To unsubscribe from this group and stop receiving ...
    Denis LamanovDenis Lamanov
    Nov 11, 2014 at 3:56 pm
    Nov 11, 2014 at 3:56 pm
  • Hi, I saw where were quite significant updates to UDA api in Impala 2.0, but I cant find any example of how to use the new framework. Maybe someone knows and can post an example of how to use fixed ...
    Jonas JarutisJonas Jarutis
    Nov 8, 2014 at 5:36 am
    Nov 8, 2014 at 5:36 am
  • Hello, As I mention, I am building a new Impala Cluster with 30 Data Nodes ( 2TB HDD each Server) . But I wonder how big Impala Cluster has been deployed. Thank you very much. Vào 15:25:13 UTC+7 Thứ ...
    Summer nguyenSummer nguyen
    Nov 7, 2014 at 2:30 pm
    Nov 7, 2014 at 2:30 pm
  • Hello All, RIght now we are discussing about having views for aggregated queries or tables. Our test results indicate that issuing a query on view is taking 2 times to the query issued on table ...
    Nov 7, 2014 at 5:30 am
    Nov 7, 2014 at 5:30 am
  • When using impala-shell for query containing Chinese character, some times the encoding is messed up, like below. However when I insert some spaces at the last part(marked in red) the query now back ...
    Binglin ChangBinglin Chang
    Nov 7, 2014 at 3:31 am
    Nov 7, 2014 at 3:31 am
  • The documentation is my area of responsibility. Yes, the separate-column approach is what comes across from a lot of customer use cases and questions, so we do address it quite a bit. I personally ...
    John RussellJohn Russell
    Nov 6, 2014 at 9:47 pm
    Nov 6, 2014 at 9:47 pm
  • Hi Jason, yes, 10k is the limit. Alex To unsubscribe from this group and stop receiving emails from it, send an email to impala-user+unsubscribe@cloudera.org.
    Alex BehmAlex Behm
    Nov 4, 2014 at 1:51 am
    Nov 4, 2014 at 1:51 am
  • Hi, It does not seem to be supported in impala 2.0. I can see https://issues.apache.org/jira/browse/TAJO-710 on that very subject. Can someone share the expected release date for this? Thanks To ...
    Stephane DrouinStephane Drouin
    Nov 4, 2014 at 1:41 am
    Nov 4, 2014 at 1:41 am
  • Just to make sure are you referring to table partitioning by range like: http://docs.oracle.com/cd/E17952_01/refman-5.5-en/partitioning-range.html Or window clauses with ranges like RANGE BETWEEN ...
    Justin EricksonJustin Erickson
    Nov 3, 2014 at 7:37 pm
    Nov 3, 2014 at 7:37 pm
  • Hi Dima, We are running CDH4 with Impala 2.0 and we experienced a similar issue when we were doing physical maintenance that implied shutting down a datanode. I'm not sure it did took as much as 30 ...
    Tony BussieresTony Bussieres
    Nov 3, 2014 at 3:22 pm
    Nov 3, 2014 at 3:22 pm
Group Navigation
period‹ prev | Nov 2014 | next ›
Group Overview
groupimpala-user @

42 users for November 2014

Philippe Marseille: 8 posts Joshua Klahr: 5 posts 吴朱华: 5 posts Bharath Vissapragada: 4 posts Dimitris Tsirogiannis: 4 posts Jonas Jarutis: 4 posts Tony Bussieres: 4 posts 刘俊: 4 posts Justin Erickson: 3 posts Kris Peeters: 3 posts Matthew Jacobs: 3 posts Nong Li: 3 posts Stephane Drouin: 3 posts Alex Behm: 2 posts Darren Lo: 2 posts Henry Robinson: 2 posts Marcel Kornacker: 2 posts Mauricio Aristizabal: 2 posts Shabnam perween: 2 posts Skye Wanderman-Milne: 2 posts
show more