sql: improve inverted index validation performance #35742
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 |
|
@lucy-zhang it's probably worth adding a benchmark so that future changes don't cause a regression in performance. Thanks! |
dt
approved these changes
May 13, 2019
To compute the expected number of keys in an inverted index, we currently generate every key and count the number of unique keys, since array elements can produce duplicate keys. This PR avoids some of these allocations by doing a simple count for every non-array JSON type. For a 25GB table of random JSON values generated by the workload, I got a ~15% speedup on the `select sum(crdb_internal.json_num_index_entries(v)) from json.j` query. Release note: None
thoszhang
force-pushed
the
faster-num-entries
branch
from
9c68d48
to
a1fc446
Compare
|
I added a benchmark for a JSON object containing mostly scalars and few arrays: |
|
bors r+ |
craig bot
pushed a commit
that referenced
this pull request
May 13, 2019
35742: sql: improve inverted index validation performance r=lucy-zhang a=lucy-zhang To compute the expected number of keys in an inverted index, we currently generate every key and count the number of unique keys, since array elements can produce duplicate keys. This PR avoids some of these allocations by doing a simple count for every non-array JSON type. For a 25GB table of random JSON values generated by the workload, I got a ~15% speedup on the `select sum(crdb_internal.json_num_index_entries(v)) from json.j` query. Release note: None Co-authored-by: Lucy Zhang <[email protected]>
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Outdated suggestions cannot be applied.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.
To compute the expected number of keys in an inverted index, we currently
generate every key and count the number of unique keys, since array elements
can produce duplicate keys. This PR avoids some of these allocations by doing a
simple count for every non-array JSON type. For a 25GB table of random JSON
values generated by the workload, I got a ~15% speedup on the
select sum(crdb_internal.json_num_index_entries(v)) from json.jquery.Release note: None