[Impala-CR](cdh5-trunk) IMPALA-2840: Don't store table location in partition location

["Marcel Kornacker (Code Review)" <ge...@cloudera.org>]

Marcel Kornacker has posted comments on this change.

Change subject: IMPALA-2840: Don't store table location in partition location
......................................................................


Patch Set 6:

(13 comments)

http://gerrit.cloudera.org:8080/#/c/2355/6/be/src/runtime/descriptors.cc
File be/src/runtime/descriptors.cc:

Line 43: std::string DecompressLocation(const impala::THdfsTable& thrift_table,
why not just make this static? (and remove the namespace)


Line 44:     const impala::THdfsPartition& thrift_partition) {
use a string* output param instead


http://gerrit.cloudera.org:8080/#/c/2355/6/be/src/runtime/descriptors.h
File be/src/runtime/descriptors.h:

Line 254:   // stripped.
that would be good. do you know how large the descriptor tables get today for something like tpc-ds? (or our concurrent benchmark?)


http://gerrit.cloudera.org:8080/#/c/2355/6/fe/src/main/java/com/cloudera/impala/catalog/HdfsPartitionLocationCompressor.java
File fe/src/main/java/com/cloudera/impala/catalog/HdfsPartitionLocationCompressor.java:

Line 23: import com.cloudera.impala.util.ListMap;
there are a few minor objectionable things in ListMap, could you include it in this patch?


Line 33:  *
extra blank line


Line 93:     private final int prefix_index;
_


Line 135:       if (0 == numPartitions) return new Pair<String,String>(s, "");
this is a small function and i can understand the code perfectly well, but this isn't the number of partitions, it's the number of partition columns


Line 142:         if (s.charAt(i) == '/') {
single line


http://gerrit.cloudera.org:8080/#/c/2355/6/fe/src/main/java/com/cloudera/impala/catalog/HdfsTable.java
File fe/src/main/java/com/cloudera/impala/catalog/HdfsTable.java:

Line 467:   public HdfsTable(TableId id, org.apache.hadoop.hive.metastore.api.Table msTbl,
move c'tor to top, below static classes.


Line 1265:     //new HdfsPartitionLocationCompressor(numClusteringCols_);
remove


http://gerrit.cloudera.org:8080/#/c/2355/6/testdata/workloads/functional-query/queries/QueryTest/alter-table.test
File testdata/workloads/functional-query/queries/QueryTest/alter-table.test:

Line 846: ---- QUERY
add summary comment what this is testing


Line 895: ---- QUERY
same here


Line 935: select i,s from alter_table_test_db.table_one_part order by i;
isn't show partitions a more direct way of ensuring that the partition locations didn't get changed? (if that's what you want to demonstrate)


-- 
To view, visit http://gerrit.cloudera.org:8080/2355
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I8c67b6ce0f83de2f5277a528a9ce67e47d638adb
Gerrit-PatchSet: 6
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Jim Apple <jb...@cloudera.com>
Gerrit-Reviewer: Dimitris Tsirogiannis <dt...@cloudera.com>
Gerrit-Reviewer: Jim Apple <jb...@cloudera.com>
Gerrit-Reviewer: Marcel Kornacker <ma...@cloudera.com>
Gerrit-Reviewer: Mostafa Mokhtar <mm...@cloudera.com>
Gerrit-Reviewer: Sailesh Mukil <sa...@cloudera.com>
Gerrit-HasComments: Yes