Skip to content
Merged
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/

package org.apache.iceberg.actions;

import org.apache.iceberg.expressions.Expression;

/**
* An action for converting the equality delete files according to a convert strategy.
* Generally used for optimizing the sizing and layout of delete files within a table.

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.

I am not sure the second sentence is accurate. It applies more to the rewrite action.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Hm, I think the first one should be enough, let me delete them.

*/
public interface ConvertDeleteFiles extends SnapshotUpdate<ConvertDeleteFiles, ConvertDeleteFiles.Result> {

@aokolnychyi aokolnychyi Oct 15, 2021

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.

Are there any other conversions we plan to support other than equality -> position? According to the doc, it seems to be the only conversion we consider.

If so, would it make sense to call it ConvertEqualityDeleteFiles? Just asking. Then we can drop the method below.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Make sense to me, the below function should be used in the original API definition which has both rewrite and converts stuff


/**
* Convert the equality deletes to the position deletes.
*
* @return this for method chaining
*/
ConvertDeleteFiles convertEqualityDeletes();

/**
* A filter for choosing the equality deletes to convert.
*
* @param expression An iceberg expression used to choose deletes.
* @return this for method chaining
*/
ConvertDeleteFiles filter(Expression expression);

/**
* The action result that contains a summary of the execution.
*/
interface Result {
/**
* Returns the count of the deletes that been converted.
*/
int convertedDeleteFilesCount();

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.

wondering if both interfaces can share the same Result interface. Can both call this removedDeleteFilesCount which also match the other method of addedDeleteFilesCount?

@aokolnychyi aokolnychyi Oct 15, 2021

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.

Not a strong opinion but have kept each action result separate so far.
I think it probably makes sense to me here too.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

so we prefer to have the result separately, right?


/**
* Returns the count of the added position delete files.
*/
int addedDeleteFilesCount();

@aokolnychyi aokolnychyi Oct 15, 2021

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.

If we think this action will be always about converting equality to position, then these methods can be a bit more specific just like their Javadoc.

For example, we can call them convertedEqualityDeleteFilesCount and addedPositionDeleteFilesCount.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Make sense to me. Will do.

}
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/

package org.apache.iceberg.actions;

import org.apache.iceberg.expressions.Expression;

/**
* An action for rewriting the position delete files according to a rewrite strategy.

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.

nit: In the doc, we are still debating whether a rewrite of equality deletes makes sense. I agree it is useful to have a generic name for this action and I like RewriteDeleteFiles. Should we ensure the Javadoc is generic too? What about dropping position in the class doc?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Instead of make API and doc generic, how about make them more specific? I mean change RewriteDeleteFiles to RewritePositionDeleteFiles.

* Generally used for optimizing the sizing and layout of delete files within a table.
*/
public interface RewriteDeleteFiles extends SnapshotUpdate<RewriteDeleteFiles, RewriteDeleteFiles.Result> {

/**
* Combine the position deletes.
*
* @return this for method chaining
*/
RewriteDeleteFiles combinePositionDeletes();

@aokolnychyi aokolnychyi Oct 15, 2021

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.

Question. There is consensus that bin-packing position deletes is probably the most useful rewrite. However, we may add a rewrite that would also reshuffle deletes, not only bin-pack them. We don't know whether that will be needed and when we will implement it but what about calling this binPackPositionDeletes similar to binPack in data compaction? That way, we can add another strategy later as combineXXX seems vague.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

I call this combineXXX because I think bin-pack is a kind of combined action. But this indeed may be vague if we have something like reshuffle in the future. Let me make this more specific.


/**
* A filter for choosing deletes to rewrite.
*
* @param expression An iceberg expression used to choose deletes.
* @return this for method chaining
*/
RewriteDeleteFiles filter(Expression expression);

/**
* The action result that contains a summary of the execution.
*/
interface Result {

@aokolnychyi aokolnychyi Oct 15, 2021

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.

I assume we will compact each partition separately so we may want to provide some file group stats in the future, just like we do in data compaction. Too early to tell whether that will be useful, though. Just something to keep in mind.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Thanks. We could always improve the API later.

/**
* Returns the count of the position deletes that been rewritten.
*/
int rewrittenDeleteFilesCount();

/**
* Returns the count of the added delete files.
*/
int addedDeleteFilesCount();
}
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,75 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/

package org.apache.iceberg.actions;

import java.util.Map;
import java.util.Set;
import org.apache.iceberg.DeleteFile;
import org.apache.iceberg.FileScanTask;
import org.apache.iceberg.Table;

public interface ConvertDeleteStrategy {

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.

might be helpful to add some javadoc. I assume this is for converting equality deletes to position deletes from the design doc?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

will do.

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.

@RussellSpitzer, we may want to rename RewriteStrategy into something data-specific since we are about to add more rewrite strategies.


/**
* Returns the name of this convert deletes strategy
*/
String name();

/**
* Returns the table being modified by this convert strategy
*/
Table table();

/**
* Returns a set of options which this convert strategy can use. This is an allowed-list and any options not
* specified here will be rejected at runtime.
*/
Set<String> validOptions();

/**
* Sets options to be used with this strategy
*/
RewriteDeleteStrategy options(Map<String, String> options);

/**
* Select the delete files to convert.
*
* @return iterable of original delete file to be converted.
*/
Iterable<DeleteFile> selectDeleteFiles();

/**
* Define how to convert the deletes.
*
* @param deleteFilesToConvert a group of files to be converted together
* @return iterable of delete files used to replace the original delete files.
*/
Iterable<DeleteFile> convertDeleteFiles(Iterable<DeleteFile> deleteFilesToConvert);

/**
* Groups delete files into lists which will be processed in a single executable unit. Each group will end up being
* committed as an independent set of changes. This creates the jobs which will eventually be run as by the underlying
* Action.
*
* @param dataFiles iterable of data files that contain the DeleteFile to be converted
* @return iterable of lists of FileScanTasks which will be processed together
*/
Iterable<Iterable<FileScanTask>> planDeleteFileGroups(Iterable<FileScanTask> dataFiles);

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.

what about returning Iterable<CombinedScanTask>?

@aokolnychyi aokolnychyi Oct 15, 2021

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.

Hm, I am not sure I get why we are accepting dataFiles here. It seems to be borrowed from the data compaction and does not match the design doc. I agree we should split delete files into group that will be processed independently but I am not sure what data files we will pass here.

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.

In case of data compaction, it is an iterable of data file lists to be processed independently.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

There is one minor difference between rewrite position delete and convert equality delete. Which is the convert action needs planning and rewrite does not. That's why here we use FileScan and use DeleteFile in the rewrite action. This has been changed back and forth.

}
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/

package org.apache.iceberg.actions;

import java.util.Map;
import java.util.Set;
import org.apache.iceberg.DeleteFile;
import org.apache.iceberg.Table;

public interface RewriteDeleteStrategy {

@jackye1995 jackye1995 Jul 19, 2021

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.

I think we also need the methods similar to RewriteStrategy such as name, validOptions

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

I didn't add validOptions because we haven't finalized what options are valid and it can be added late. But I'm fine to add them now.

For name, It looks like the class name already includes the strategy name. Do we still need it?

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.

name is mostly used for logging purpose. I agree it is already in the class name, but for consistency with RewriteStrategy I think we can still have it here.

/**
* Returns the name of this rewrite deletes strategy
*/
String name();

/**
* Returns the table being modified by this rewrite strategy
*/
Table table();

/**
* Returns a set of options which this rewrite strategy can use. This is an allowed-list and any options not
* specified here will be rejected at runtime.
*/
Set<String> validOptions();

/**
* Sets options to be used with this strategy
*/
RewriteDeleteStrategy options(Map<String, String> options);

/**
* Select the delete files to rewrite.
*
* @return iterable of original delete file to be replaced.
*/
Iterable<DeleteFile> selectDeleteFiles();

/**
* Define how to rewrite the deletes.
*
* @param deleteFilesToRewrite a group of files to be rewritten together
* @return iterable of delete files used to replace the original delete files.
*/
Iterable<DeleteFile> rewriteDeleteFiles(Iterable<DeleteFile> deleteFilesToRewrite);

/**
* Groups into lists which will be processed in a single executable unit. Each group will end up being
* committed as an independent set of changes. This creates the jobs which will eventually be run as by the underlying
* Action.
*
* @param deleteFiles iterable of DeleteFile to be rewritten
* @return iterable of lists of FileScanTasks which will be processed together
*/
Iterable<Iterable<DeleteFile>> planDeleteFileGroups(Iterable<DeleteFile> deleteFiles);

@stevenzwu stevenzwu Oct 15, 2021

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.

wondering why here it is DeleteFile while the ConvertDeleteStrategy plans with FileScanTask.

also wondering if both interfaces in this PR can use RewriteFileGroup as return type? @RussellSpitzer

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

There is one difference between the rewrite position delete and the convert, it is about the planning.

}