https://community.cloudera.com/t5/Support-Questions/Count-Distinct-discrepancy-Hive/m-p/225085#M186948 <P><A href="https://community.cloudera.com/legacyfs/online/attachments/77763-screen-shot-2018-06-19-at-104245-pm.png">screen-shot-2018-06-19-at-104245-pm.png</A></P><P><A href="https://community.cloudera.com/legacyfs/online/attachments/77763-screen-shot-2018-06-19-at-104245-pm.png"></A>Count distinct doesn't always give me the right answer. I've attached two different queries that should both result in 7 unique items purchased. If I don't do an operation on mdse_item_i like cast it to a bigint, it doesn't always count them correctly.</P><P>to explain it simply, when i do cast on "mdse_it<A href="https://community.cloudera.com/legacyfs/online/attachments/77760-countsdistinct-1.txt">countsdistinct-1.txt</A><A href="https://community.cloudera.com/legacyfs/online/attachments/77761-doesnotcountdistinct.txt">doesnotcountdistinct.txt</A>em_i" it gives unique results as 7, but when i don't do casting it gives unique results as 10 which is not correct.</P><P>hive&gt; select * from dfr_distinct; OK 100000000938 5<STRONG> 7</STRONG> 12.33 2 2.75 4.27 8.060 2 8 0 Time taken: 0.479 seconds, Fetched: 1 row(s) hive&gt; select * from dfr_distinctnot; OK 100000000938 5 <STRONG>10</STRONG> 12.33 2 2.75 4.27 8.06 0 2 8 0 Time taken: 0.932 seconds, Fetched: 1 row(s)</P><P>tried running the query in both MR and Tez modes still giving same results when i don't do casting.</P> Wed, 20 Jun 2018 10:41:58 GMT reddyr211 2018-06-20T10:41:58Z