----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/24498/#review50938 -----------------------------------------------------------
metastore/src/java/org/apache/hadoop/hive/metastore/IExtrapolatePartStatus.java <https://reviews.apache.org/r/24498/#comment88831> Needs apache license header. Look at top of any other java file. metastore/src/java/org/apache/hadoop/hive/metastore/IExtrapolatePartStatus.java <https://reviews.apache.org/r/24498/#comment88832> Better name: AggrType? metastore/src/java/org/apache/hadoop/hive/metastore/LinearExtrapolatePartStatus.java <https://reviews.apache.org/r/24498/#comment88835> Apache header? metastore/src/java/org/apache/hadoop/hive/metastore/MetaStoreDirectSql.java <https://reviews.apache.org/r/24498/#comment88821> It will be good if this api also returns count for # of partitions for which stats were found. metastore/src/java/org/apache/hadoop/hive/metastore/MetaStoreDirectSql.java <https://reviews.apache.org/r/24498/#comment88822> I think checking total == (# of cols) * (# of parts) is better. metastore/src/java/org/apache/hadoop/hive/metastore/MetaStoreDirectSql.java <https://reviews.apache.org/r/24498/#comment88829> We are doing 3 queries per column. We need to do better here. e.g if there are 20 columns in table. We will end up with 60 queries. Tables with few hundred columns are not unheard of. Seems like query for column type can be avoided altogether, since column names are sent from client, we can also send type info from client, since client already has it. For other queries, also we need to make it independent of # of columns. ql/src/test/queries/clientpositive/extrapolate_part_stats.q <https://reviews.apache.org/r/24498/#comment88830> Why do you have this flag set to false? Unless, there is a reason take this off. - Ashutosh Chauhan On Aug. 17, 2014, 4:23 a.m., pengcheng xiong wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/24498/ > ----------------------------------------------------------- > > (Updated Aug. 17, 2014, 4:23 a.m.) > > > Review request for hive. > > > Repository: hive-git > > > Description > ------- > > We propose a method to extrapolate the missing column status for the > partitions. > > > Diffs > ----- > > > metastore/src/java/org/apache/hadoop/hive/metastore/IExtrapolatePartStatus.java > PRE-CREATION > > metastore/src/java/org/apache/hadoop/hive/metastore/LinearExtrapolatePartStatus.java > PRE-CREATION > metastore/src/java/org/apache/hadoop/hive/metastore/MetaStoreDirectSql.java > 767cffc > ql/src/test/queries/clientpositive/extrapolate_part_stats.q PRE-CREATION > ql/src/test/results/clientpositive/extrapolate_part_stats.q.out > PRE-CREATION > > Diff: https://reviews.apache.org/r/24498/diff/ > > > Testing > ------- > > > File Attachments > ---------------- > > HIVE-7654.0.patch > > https://reviews.apache.org/media/uploaded/files/2014/08/12/77b155b0-a417-4225-b6b7-4c8c6ce2b97d__HIVE-7654.0.patch > > > Thanks, > > pengcheng xiong > >
