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
Full ∞ support for numeric context #97
Conversation
@t-kataym , I have updated this branch and resolved file conflict. Thank for #78 merging! PostgreSQL 13 testing problems is specific for Ubuntu and if I'll correct Update: |
@mkgrgis Thank you for the update. We will confirm it. |
Hello @mkgrgis, thanks for your hard work.
I don't understand your purpose of this description. |
Hello, @MinhLA1410 !
In PostgreSQL ±∞ values sorted before or after any usual numeric values according ISO:SQL, in SQLite infinity values with
No. This test was only unified with the file from 16.0 version.
Yes. I meant no other changes to |
Thanks @mkgrgis !
Is that correct? |
Yes, absolutely correct. |
Thanks @mkgrgis ,
This specification is not yet clear and is under ongoing discussion. Most of the source code test code related to it has been commented out. So could you separate them into another PR? (We would like to keep the code in master branch clean, with only official code (and comment) that has been tested and verified.) |
Will the separate PR "Add comments about
|
I understand it.
But as your comment above, the read/write/filter capability of We cannot accept these now. We don't accept TODO comments in the source code, only official code + comments can merge to the master |
Ok, @MinhLA1410 . I'll remove this TCs and code fragments. |
@MinhLA1410 , will 1db7f67 enough? |
@mkgrgis ,
It's OK. Thank you.
I would like to confirm this point again. As you said, I can understand that Before your implementation, sqlite_fdw cannot sorted values correctly (mean -infinity < - 1 < 0 < 1 < +infinity), after your implementation sqlite_fdw can sorted values correctly.
If I change column f of table "type_FLOAT_INF" to text type (case text affinity) and use your branch to sort. the results on postgres are same on sqlite server.
I don't see the match between current behavior with your description above. Could you explain it more? Have I misunderstood something? |
This is for case where all values have |
@mkgrgis ,
If the table mixed real affinity and text affinity of ∞ value forms. SQLite cannot sort correctly in arithmetic context. |
Yes. Because in PostgreSQL there is 2 possible methods of ∞ value input: text constant like |
@MinhLA1410 , just for info around of this PR and about different ∞ processing between PostgreSQL and SQLite . From https://www.sqlite.org/releaselog/3_45_3.html
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@mkgrgis Thanks for your support.
I reviewed you PR. Could you check my comments?
All review rounds are completed, all comments are checked, @MinhLA1410. Does this means the next round will de done by @t-kataym ? |
@MinhLA1410 Thank you for reviewing. |
@mkgrgis Thank you for your contribution. This PR was merged. |
Thanks, to pgspider team, @t-kataym ! Now I can prepare PR about |
In this PR:
9e999
for SQLite+Infinity
,-inf
,Inf
etc.float
/numeric
values in many tests from direct reading to data normalization function.NaN
values, add detailed comments about current processing ofNaN
values, but no changes to currentNaN
processing. Discussing aboutNaN
will be continues in NaN values converted to 0 #36.aggregates
test with version of test file for PostgreSQL 16.0.