Skip to content

[enhancement](csv_reader)Optimize the reading efficiency of nullable (string) columns. - #24698

Merged
morningman merged 4 commits into
apache:masterfrom
hubgeter:opt_stream_load
Sep 22, 2023
Merged

[enhancement](csv_reader)Optimize the reading efficiency of nullable (string) columns.#24698
morningman merged 4 commits into
apache:masterfrom
hubgeter:opt_stream_load

Conversation

@hubgeter

@hubgeterhubgeter commented Sep 20, 2023•
edited
Loading

Copy link
Copy Markdown
Contributor

Proposed changes

Issue Number: close #xxx
Optimize the performance of stream load tsv by reducing virtual function calls .
(Optimize read performance of nullable (string) columns by reducing virtual function calls.)
before : 600+ s
after : 560+ s

Further comments

If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...

@hubgeter
hubgeter marked this pull request as ready for review September 20, 2023 12:28
@github-actions

github-actionsBot commented Sep 20, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@hubgeter

hubgeter commented Sep 20, 2023

Copy link
Copy Markdown
ContributorAuthor

run clickbench

@doris-robot

doris-robot commented Sep 20, 2023

Copy link
Copy Markdown

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 44.85 seconds
stream load tsv: 611 seconds loaded 74807831229 Bytes, about 116 MB/s
stream load json: 21 seconds loaded 2358488459 Bytes, about 107 MB/s
stream load orc: 64 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 28.7 seconds inserted 10000000 Rows, about 348K ops/s
storage size: 17162321648 Bytes

@hubgeter

hubgeter commented Sep 20, 2023

Copy link
Copy Markdown
ContributorAuthor

run clickbench

@github-actions

github-actionsBot commented Sep 20, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@doris-robot

doris-robot commented Sep 20, 2023

Copy link
Copy Markdown

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 45.36 seconds
stream load tsv: 594 seconds loaded 74807831229 Bytes, about 120 MB/s
stream load json: 20 seconds loaded 2358488459 Bytes, about 112 MB/s
stream load orc: 64 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 29.2 seconds inserted 10000000 Rows, about 342K ops/s
storage size: 17162388002 Bytes

@morningman

morningman commented Sep 21, 2023

Copy link
Copy Markdown
Contributor

run buildall

@doris-robot

doris-robot commented Sep 21, 2023

Copy link
Copy Markdown

TeamCity be ut coverage result:
Function Coverage: 36.40% (8016/22025)
Line Coverage: 28.68% (64350/224347)
Region Coverage: 27.58% (33409/121152)
Branch Coverage: 24.24% (17097/70544)
Coverage Report: http://coverage.selectdb-in.cc/coverage/318a4bb17dff2951e529e9c1de34353e99c20e5e_318a4bb17dff2951e529e9c1de34353e99c20e5e/report/index.html

@doris-robot

doris-robot commented Sep 21, 2023

Copy link
Copy Markdown

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 45.86 seconds
stream load tsv: 594 seconds loaded 74807831229 Bytes, about 120 MB/s
stream load json: 21 seconds loaded 2358488459 Bytes, about 107 MB/s
stream load orc: 65 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 29.1 seconds inserted 10000000 Rows, about 343K ops/s
storage size: 17162331852 Bytes

morningman
morningman reviewed Sep 21, 2023
View reviewed changes
Comment threadbe/src/vec/exec/format/csv/csv_reader.cpp Outdated
} else {
_options.map_key_delim = _params.file_attributes.text_params.mapkv_delimiter[0];
}
if (is_load) {

@morningmanmorningmanSep 21, 2023

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Better check again, check if all columns are nullable string, in case that FE planner may change the behavior

@hubgeter

hubgeter commented Sep 21, 2023

Copy link
Copy Markdown
ContributorAuthor

run buildall

@github-actions

github-actionsBot commented Sep 21, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@hubgeter

hubgeter commented Sep 21, 2023

Copy link
Copy Markdown
ContributorAuthor

run buildall

@github-actions

github-actionsBot commented Sep 21, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@github-actions

github-actionsBot commented Sep 21, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@hubgeterhubgeter changed the title [enhancement](serde)enhancement serde performance.[enhancement](csv_reader)Optimize the reading efficiency of nullable (string) columns.Sep 21, 2023
@hubgeter

hubgeter commented Sep 22, 2023

Copy link
Copy Markdown
ContributorAuthor

run buildall

@doris-robot

doris-robot commented Sep 22, 2023

Copy link
Copy Markdown

TeamCity be ut coverage result:
Function Coverage: 36.41% (8023/22037)
Line Coverage: 28.70% (64446/224562)
Region Coverage: 27.61% (33475/121242)
Branch Coverage: 24.27% (17129/70564)
Coverage Report: http://coverage.selectdb-in.cc/coverage/b91a9afb096ce1bcdfeb63b397cc89cc21d9d74e_b91a9afb096ce1bcdfeb63b397cc89cc21d9d74e/report/index.html

@doris-robot

doris-robot commented Sep 22, 2023

Copy link
Copy Markdown

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 45.9 seconds
stream load tsv: 591 seconds loaded 74807831229 Bytes, about 120 MB/s
stream load json: 20 seconds loaded 2358488459 Bytes, about 112 MB/s
stream load orc: 64 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 28.9 seconds inserted 10000000 Rows, about 346K ops/s
storage size: 17162385506 Bytes

HappenLee
HappenLee approved these changes Sep 22, 2023
View reviewed changes

@HappenLeeHappenLee left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@github-actionsgithub-actionsBot added the approved Indicates a PR has been approved by one committer. label Sep 22, 2023
@github-actions

github-actionsBot commented Sep 22, 2023

Copy link
Copy Markdown
Contributor

PR approved by at least one committer and no changes requested.

@github-actions

github-actionsBot commented Sep 22, 2023

Copy link
Copy Markdown
Contributor

PR approved by anyone and no changes requested.

@github-actions

github-actionsBot commented Sep 22, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@hubgeter

hubgeter commented Sep 22, 2023

Copy link
Copy Markdown
ContributorAuthor

run buildall

@github-actions

github-actionsBot commented Sep 22, 2023

Copy link
Copy Markdown
Contributor

clang-tidy review says "All clean, LGTM! 👍"

@doris-robot

doris-robot commented Sep 22, 2023

Copy link
Copy Markdown

TeamCity be ut coverage result:
Function Coverage: 36.39% (8021/22039)
Line Coverage: 28.68% (64418/224592)
Region Coverage: 27.58% (33448/121257)
Branch Coverage: 24.25% (17117/70572)
Coverage Report: http://coverage.selectdb-in.cc/coverage/bd7df1ad0b3e7b0ccb9604a7d16ba245e161a3cd_bd7df1ad0b3e7b0ccb9604a7d16ba245e161a3cd/report/index.html

@doris-robot

doris-robot commented Sep 22, 2023

Copy link
Copy Markdown

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 45.54 seconds
stream load tsv: 562 seconds loaded 74807831229 Bytes, about 126 MB/s
stream load json: 20 seconds loaded 2358488459 Bytes, about 112 MB/s
stream load orc: 65 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 28.6 seconds inserted 10000000 Rows, about 349K ops/s
storage size: 17161970118 Bytes

@morningman
morningman merged commit 3c99743 into apache:masterSep 22, 2023
xiaokang pushed a commit that referenced this pull request Sep 22, 2023
@hubgeter@xiaokang
[enhancement](csv_reader)Optimize the reading efficiency of nullable …
0da08d5
…(string) columns. (#24698)
Optimize the performance of stream load tsv by reducing virtual function calls .
(Optimize read performance of nullable (string) columns by reducing virtual function calls.)
before : 600+ s
after : 560+ s
@xiaokangxiaokang mentioned this pull request Sep 28, 2023
Closed
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@morningmanmorningmanmorningman approved these changes
@HappenLeeHappenLeeHappenLee approved these changes

Assignees

No one assigned

Labels

approvedIndicates a PR has been approved by one committer.dev/2.0.2-mergedp0_breviewed

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@hubgeter@doris-robot@morningman@HappenLee@xiaokang