Skip to content
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.

Sign up for GitHub

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

Jump to bottom

HPCC-33155 Revisit Parquet Test Suite #19405

Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

HPCC-33155 Revisit Parquet Test Suite #19405

wants to merge 3 commits into from

Conversation

jackdelv
Copy link
Contributor

Change the plugin to throw an error if a file is not found when opening a file for reading

Change output filename of partitioned file parts. part_00_0.parquet -> part_0_of_table_0_from_worker_0.parquet
These files aren't meant to be read directly, but I think the new format makes more sense when trying to debug.

Update tests so that they are all being run and are able to use external files in testing/regress/download

Remove parquetString test and key file and move testing to parquetTypes.ecl

Update test cases in parquetTypes.ecl

Type of change:

  • This change is a bug fix (non-breaking change which fixes an issue).
  • This change is a new feature (non-breaking change which adds functionality).
  • This change improves the code (refactor or other change that does not change the functionality)
  • This change fixes warnings (the fix does not alter the functionality or the generated code)
  • This change is a breaking change (fix or feature that will cause existing behavior to change).
  • This change alters the query API (existing queries will have to be recompiled)

Checklist:

  • My code follows the code style of this project.
    • My code does not create any new warnings from compiler, build system, or lint.
  • The commit message is properly formatted and free of typos.
    • The commit message title makes sense in a changelog, by itself.
    • The commit is signed.
  • My change requires a change to the documentation.
    • I have updated the documentation accordingly, or...
    • I have created a JIRA ticket to update the documentation.
    • Any new interfaces or exported functions are appropriately commented.
  • I have read the CONTRIBUTORS document.
  • The change has been fully tested:
    • I have added tests to cover my changes.
    • All new and existing tests passed.
    • I have checked that this change does not introduce memory leaks.
    • I have used Valgrind or similar tools to check for potential issues.
  • I have given due consideration to all of the following potential concerns:
    • Scalability
    • Performance
    • Security
    • Thread-safety
    • Cloud-compatibility
    • Premature optimization
    • Existing deployed queries will not be broken
    • This change fixes the problem, not just the symptom
    • The target branch of this pull request is appropriate for such a change.
  • There are no similar instances of the same problem that should be addressed
    • I have addressed them here
    • I have raised JIRA issues to address them separately
  • This is a user interface / front-end modification
    • I have tested my changes in multiple modern browsers
    • The component(s) render as expected

Smoketest:

  • Send notifications about my Pull Request position in Smoketest queue.
  • Test my draft Pull Request.

Testing:

@jackdelv jackdelv requested a review from dcamper January 13, 2025 15:39
@github-actions GitHub Actions
Copy link

Jira Issue: https://hpccsystems.atlassian.net//browse/HPCC-33155

Jirabot Action Result:
Workflow Transition To: Merge Pending
Updated PR

dcamper
dcamper reviewed Jan 13, 2025
View reviewed changes
Copy link
Contributor

@dcamper dcamper left a comment

Choose a reason for hiding this comment

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

Just one question about the contents of a comment block.

testing/regress/ecl/parquetPartition.ecl
* ├── district=1
* │ ├── firstname=Alice
* │ │ └── city=New%20York
* │ │ └── part_0_of_table_0_from_worker_0.parquet
Copy link
Contributor

Choose a reason for hiding this comment

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

These part... names are files? If so, they are all the same for every value combination. Is that correct?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, that is correct.

@dcamper dcamper self-requested a review January 13, 2025 16:01
dcamper
dcamper approved these changes Jan 13, 2025
View reviewed changes
Copy link
Contributor

@dcamper dcamper left a comment

Choose a reason for hiding this comment

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

Looks good to me.

ghalliday
ghalliday reviewed Jan 14, 2025
View reviewed changes
Copy link
Member

@ghalliday ghalliday left a comment

Choose a reason for hiding this comment

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

See comments/questions.

testing/regress/ecl/parquetCorrupt.ecl
DATASET([{0, 'Corrupt Parquet File', ''}], RECORDDEF),
CORRUPT_PARQUET);

OUTPUT(CORRUPT_RESULT);
Copy link
Member

Choose a reason for hiding this comment

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

Without the output this workunit will not do anything.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Fixed.

testing/regress/ecl/parquetEmpty.ecl
@@ -31,6 +36,3 @@ EMPTY_PARQUET := DATASET([], RECORDDEF);
ParquetIO.Write(EMPTY_PARQUET, filePath, TRUE);

read_data := ParquetIO.Read(RECORDDEF, filePath);

OUTPUT(read_data);
Copy link
Member

Choose a reason for hiding this comment

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

With this deleted it will not read the file.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Fixed.

@jackdelv
Add a sort to vector of FileReaders
111a0d2 
- createDirectoryIterator is used to list all matching Parquet files and the order is not guarenteed
- Fix parquetEmpty.ecl and parquetCorrupt.ecl by adding the OUTPUT statements back
@jackdelv jackdelv requested review from ghalliday and dcamper January 15, 2025 19:01
dcamper
dcamper approved these changes Jan 15, 2025
View reviewed changes
Copy link
Contributor

@dcamper dcamper left a comment

Choose a reason for hiding this comment

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

Approved.

ghalliday
ghalliday reviewed Jan 16, 2025
View reviewed changes
plugins/parquet/parquetembed.cpp
@@ -430,7 +441,7 @@ arrow::Result<std::shared_ptr<arrow::Table>> ParquetReader::queryRows()
{
// Start by getting the number of rows in the first group and checking if it includes this workers startRow
__int64 offset = (*rbatchItr)->get()->num_rows();
while (offset < startRow)
Copy link
Member

Choose a reason for hiding this comment

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

This looks like a bug fix - is it?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, I think the changes to the openReadFile would be considered a bug fix as well. Should they be pulled out into separate PRs?

ghalliday
ghalliday reviewed Jan 16, 2025
View reviewed changes
Copy link
Member

@ghalliday ghalliday left a comment

Choose a reason for hiding this comment

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

Two questions.

plugins/parquet/parquetembed.cpp
@@ -381,7 +392,7 @@ arrow::Status ParquetReader::processReadFile()
rbatchItr = arrow::RecordBatchReader::RecordBatchReaderIterator(rbatchReader.get());
PARQUET_ASSIGN_OR_THROW(auto datasetRows, scanner->CountRows());
// Divide the work among any number of workers
divide_row_groups(activityCtx, datasetRows, totalRowCount, startRowGroup);
divide_row_groups(activityCtx, datasetRows, totalRowCount, startRow);
Copy link
Member

Choose a reason for hiding this comment

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

Is this change deliberate. If so, should line 409 also change

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, that was deliberate. startRow is used for arrow partitioned datasets because the API returns rows rather than groups of rows. This was an error causing multi node thor workers to always start reading from the first row.

@jackdelv jackdelv requested a review from ghalliday January 16, 2025 16:34
@ghalliday
Copy link
Member

@jackdelv how important are the bug fixes? Should they be targeting an earlier build than 9.10.x?
I don't mind merging as-is if we consider parquet as something that users are unlikely to be using at the moment. If we consider it possibly in use, then they should be separated out.

@AttilaVamos
Copy link
Contributor

@jackdelv I raised a JIRA, the HPCC-33278 related to Test Suite.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@dcamper dcamper dcamper approved these changes

@ghalliday ghalliday Awaiting requested review from ghalliday

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@jackdelv @ghalliday @AttilaVamos @dcamper