[IMPALA-15] UNIX_TIMESTAMP format behaviour deviates from Hive when format matches a prefix of the time value

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Impala 0.3
• Assignee: Skye Wanderman-Milne
• Reporter: Henry Robinson
• Created: 08/Jan/13 23:20
• Updated: 28/Jan/13 19:56
• Resolved: 28/Jan/13 19:56

Description

Impala's UNIX_TIMESTAMP(val, format) operation compares the length of format and val and returns NULL if they do not match. Hive instead effectively truncates val to format's length.

For example, in Hive:

hive> select UNIX_TIMESTAMP('2000-02-01 10:20:30','yyyy-MM') FROM tmp;
<snip>
OK
949392000

hive> select UNIX_TIMESTAMP('2000-02-01 10:20:30','yyyy') FROM tmp;
<snip>
OK
946713600 <<< Note that value has changed since the format truncated the string to just '2000'

hive> select UNIX_TIMESTAMP('2000-02-01 10:20:30 junk text','yyyy') FROM tmp;
<snip>
OK
946713600 <<< Still works! Garbage string is ignored.