Apache Hive

Compute statistics for partitions

1. Get a list of partitions into a file:
   hdfs dfs -ls /hive/dabase/path/table/ >partitions.txt
   Leave only partition names in the file,remove paths and dates
2. Generate statements
   cat partitions.txt | while read line
   do
   echo "analyze mydb.mytable partition(${line}) compute statistics;
   done >stats.sql
3. Run the resultant script in hive