impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-4549: consistently treat 9999 as upper bound for timestamp year
Date Wed, 11 Jan 2017 00:35:49 GMT
Tim Armstrong has posted comments on this change.

Change subject: IMPALA-4549: consistently treat 9999 as upper bound for timestamp year

Patch Set 1:

Commit Message:

Line 9: Use the patched boost version and update tests accordingly.
> Maybe briefly describe the behavior of Impala before this patch?
File be/src/exprs/

Line 1482:   TestIsNull("cast(cast('9999-12-31 23:59:59' as timestamp) + interval 1 year as
> I think this test would be more interesting if we add 1 second instead. (th
I looked at this again and realised I misunderstood what it was doing.

I think this was a workaround to deal with the fact that we couldn't parse 5-digit years 10000.
We have more extensive edge case tests below in the timestamp tests. I modified one of the
below tests to test + 1 second.

Line 1484:   TestValue("cast(cast('9999-12-31 23:59:59' as timestamp) as bigint)",
> This seems like the base case, put it above the TestIsNull case. (maybe add

Line 1486:   TestTimestampValue("cast(253402300799 as timestamp) - interval 1 year",
> Not sure what this is testing. Maybe remove?
It's testing the int->timestamp reverse conversion. I restructured this to be a bit more
logically organised.

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: Iaf23b40833017789d879e5da7bb10384129e2d10
Gerrit-PatchSet: 1
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Tim Armstrong <>
Gerrit-Reviewer: Taras Bobrovytsky <>
Gerrit-Reviewer: Tim Armstrong <>
Gerrit-HasComments: Yes

View raw message