[Hadoop-common-user] Reading a common library file

I need to read and check the entry from a common file ( say a list of words
containing StopWords ) . So that each map function can check these stopwords
from a file.

Should I create a hadoop map file in HDFS so that it could be searched
faster ? would that be memory overhead as indexes would be loaded into
memory ?

another approach could be a Grep in map/.reduce job

Any suggestions would be appreciated

thanks