-
Notifications
You must be signed in to change notification settings - Fork 348
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Always map java.sql.Types.BIGINT to LongType #311
base: master
Are you sure you want to change the base?
Conversation
java.sql.Types.BIGINT => LongType in all cases for Redshift. Redshift has no unsigned BigInt, and thus BigInt maps to Long without loss of resolution.
Thanks for the fix! Mind adding a regression test to an existing suite? If
you can't, I can tackle it in a followup. Code from your original issue
would be a good foundation for the test case. Check out RedshiftReadSuite,
since that's where I'd add it.
…On Fri, Dec 9, 2016 at 1:20 AM ngerstle ***@***.***> wrote:
java.sql.Types.BIGINT => LongType in all cases for Redshift.
Redshift has no unsigned BigInt, and thus BigInt maps to Long without loss
of resolution.
See issue #310 <#310>
------------------------------
You can view, comment on, or merge this pull request online at:
#311
Commit Summary
- Update RedshiftJDBCWrapper.scala
File Changes
- *M*
src/main/scala/com/databricks/spark/redshift/RedshiftJDBCWrapper.scala
<https://github.com/databricks/spark-redshift/pull/311/files#diff-0>
(2)
Patch Links:
- https://github.com/databricks/spark-redshift/pull/311.patch
- https://github.com/databricks/spark-redshift/pull/311.diff
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#311>, or mute the
thread
<https://github.com/notifications/unsubscribe-auth/AADGPO5Ox0OGiqp5pwDLug4_usBaAuAbks5rGR1AgaJpZM4LIulr>
.
|
Errors/unusually behavior is associated with BigInts. Specifically, a count(column) query sometimes returns doubles rather than longs, as would be expected from the count function (which only provides signed BigInts).
I'm not able to find any clear point at which a decimal would be converted to a double, meaning that issue #310 is unlikely to fixed by this, but the dead code should still be pruned. |
Current coverage is 88.55% (diff: 100%)@@ master #311 diff @@
==========================================
Files 15 15
Lines 769 769
Methods 614 615 +1
Messages 0 0
Branches 155 154 -1
==========================================
Hits 681 681
Misses 88 88
Partials 0 0
|
Regarding the double issue, is there a possibility that our call to ask Redshift the schema is returning a Double but then Redshift is actually turning around and unexpectedly giving us a Decimal? Do you have any test case which reproduces the Double behavior, even if you don't know how to actually fix the bug? |
Unfortunately, I'm still trying to get a consistent abstract case- as mentioned in #310, I have a |
java.sql.Types.BIGINT => LongType in all cases for Redshift.
Redshift has no unsigned BigInt, and thus BigInt maps to Long without loss of resolution.
See issue #310