[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user ArvindSridhar commented on the issue: https://github.com/apache/madlib/pull/292 PR ready to be closed, changes made elsewhere already ---

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/292 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/559/ ---

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/292 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/558/ ---

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user ArvindSridhar commented on the issue: https://github.com/apache/madlib/pull/292 Sounds good. Just pushed a commit to this PR, will rebase and add back the original function to the original vec2cols PR tomorrow ---

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user iyerr3 commented on the issue: https://github.com/apache/madlib/pull/292 Yeah, the unit tests are unnecessary after this change. I vote to remove them, with just the function added back to the original PR. On Thu, Jul 12, 2018 at 5:26 PM Arvind Sridhar

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user ArvindSridhar commented on the issue: https://github.com/apache/madlib/pull/292 Sounds good-just verified that plpy converts the SQL bool into a Python bool and not a Python string. It should be good to go now. Do we still want to get rid of the unit tests now given that t

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user iyerr3 commented on the issue: https://github.com/apache/madlib/pull/292 That's a valid concern. You could return bool(result[0]['n_y']) to ensure a bool output instead of a string (though I believe the SPI ensures that boolean from SQL is returned as bool in Python)

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user ArvindSridhar commented on the issue: https://github.com/apache/madlib/pull/292 I changed the query but was wondering: since we are now returning just result[0]['n_y'] in python, could this cause problems if our functions calling this expect a python boolean output, while

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user iyerr3 commented on the issue: https://github.com/apache/madlib/pull/292 Few comments: 1. The query would read better if it is setup to return a boolean i.e. using `array_upper(..., 2) IS NULL` instead of converting the integer output to bool in Python. 2. My req

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/292 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/554/ ---

[GitHub] madlib issue #292: Internal: Add function to check column type for 1D array

Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/292 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/553/ ---