builtins: add extract functionality for intervals #43293
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This change is |
otan
force-pushed
the
extract_interval
branch
from
73b9767
to
17a344f
Compare
knz
approved these changes
Jan 8, 2020
There was a problem hiding this comment.
Very nice
However please change the release note to add an actual deprecation notice for extract_duration, as well as a note to the same effect in that function's text description.
Thanks
Reviewed 4 of 4 files at r1.
Reviewable status:complete! 1 of 0 LGTMs obtained (waiting on @solongordon)
This adds the postgres-esque implementation of extract for intervals. Tested a variety of edge cases and compared against what postgres returns. Currently leaving the `extract_duration` function alone - we could thereoetically get rid of it after this PR. It does behave slightly differently as well. Release note (sql change): Added new ability to call extract on an interval, e.g. `extract(day from interval '254 days')`. This follows the postgres implementation. Furthermore, this deprecates `extract_duration`, which will be removed at a later date.
|
bors r=knz thanks for the review :) |
craig bot
pushed a commit
that referenced
this pull request
Jan 9, 2020
39308: sql/sem: fix interaction between framing and peers in window functions r=yuzefovich a=yuzefovich **rowexec: minor refactor of partitionPeerGrouper** This commit removes allPeers peerGroupChecker and refactors partitionPeerGrouper to support the case of omitted ORDER BY clause. Release note: None **sql/sem: fix interaction between framing and peers in window functions** Previously, whenever we had a result of a window function already computed for a peer of the row, we would return that result. However, the concept of window framing takes precedence over the concept of peers (i.e. if the window frame over which we computed the result for a peer is different from the window frame of the current row, we need to recompute the result). Now this is fixed by tracking the boundaries of the window frame over which peerRes is computed and recomputing the latter when needed. There was also a bug in IsDefaultFrame where we forgot to check whether the mode of framing is RANGE. This caused incorrect results in some cases when ROWS or GROUPS modes of framing were used. Fixes: #38901. Fixes: #42935. Release note (bug fix): CockroachDB previously was returning incorrect results for some aggregate functions when used as window functions with non-default window frame. Now this is fixed. Note that MIN, MAX, SUM, AVG as well as "pure" window functions (i.e. non-aggregates) were not affected. 43293: builtins: add extract functionality for intervals r=knz a=otan Resolves #43209 This adds the postgres-esque implementation of extract for intervals. Tested a variety of edge cases and compared against what postgres returns. Currently leaving the `extract_duration` function alone - we could thereoetically get rid of it after this PR. It does behave slightly differently as well. Release note (sql change): Added new ability to call extract on an interval, e.g. `extract(day from interval '254 days')`. This follows the postgres implementation. Furthermore, this deprecates `extract_duration`, which will be removed at a later date. 43834: colexec: speed up ordered sync r=yuzefovich a=yuzefovich **colexec: refactor ordered sync to use assign instead of append** Ordered synchronizer sets the output one row at a time. Previously, we were calling Append on the output vectors, but it is faster to use Sets in such scenario. Release note: None **colexec: speed up ordered sync** This commit speeds up the ordered synchronizer with two modifications: 1. instead of updating the allocator after we copy a single whole row, we will now update the allocator after we copy a full batch of rows 2. to reduce the number of interface conversions we will "strip" output vectors to get to the underlying slices and will use those directly when copying the data. Release note: None Co-authored-by: Yahor Yuzefovich <yahor@cockroachlabs.com> Co-authored-by: Oliver Tan <otan@cockroachlabs.com>
Build succeeded |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Resolves #43209
This adds the postgres-esque implementation of extract for intervals.
Tested a variety of edge cases and compared against what postgres
returns.
Currently leaving the
extract_durationfunction alone - we couldthereoetically get rid of it after this PR. It does behave slightly
differently as well.
Release note (sql change): Added new ability to call extract on an
interval, e.g.
extract(day from interval '254 days'). This followsthe postgres implementation. Furthermore, this deprecates
extract_duration, which will be removed at a later date.