You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by "manuzhang (via GitHub)" <gi...@apache.org> on 2023/05/19 05:19:27 UTC
[GitHub] [iceberg] manuzhang opened a new pull request, #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
manuzhang opened a new pull request, #7653:
URL: https://github.com/apache/iceberg/pull/7653
Back-port of https://github.com/apache/iceberg/pull/4652 to Spark 3.1
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho merged pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho merged PR #7653:
URL: https://github.com/apache/iceberg/pull/7653
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1266249371
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
+ row -> {
+ FileMetadata actual = row._1;
+ FileMetadata valid = row._2;
+
+ if (valid == null) {
+ return actual.location;
+ }
+
+ boolean schemeMatch =
+ Strings.isNullOrEmpty(valid.scheme)
+ || valid.scheme.equalsIgnoreCase(actual.scheme);
+ boolean authorityMatch =
+ Strings.isNullOrEmpty(valid.authority)
+ || valid.authority.equalsIgnoreCase(actual.authority);
+
+ if ((!schemeMatch || !authorityMatch) && mode == PrefixMismatchMode.DELETE) {
+ return actual.location;
+ } else {
+ if (!schemeMatch) {
+ conflicts.add(Pair.of(valid.scheme, actual.scheme));
+ }
+ if (!authorityMatch) {
+ conflicts.add(Pair.of(valid.authority, actual.authority));
+ }
+ }
+
+ return null;
+ });
+ return Iterators.filter(transformed, Objects::nonNull);
+ };
+ }
+
+ private static Map<String, String> flattenMap(Map<String, String> map) {
+ Map<String, String> flattenedMap = Maps.newHashMap();
+ if (map != null) {
+ for (String key : map.keySet()) {
+ String value = map.get(key);
+ for (String splitKey : COMMA.split(key)) {
+ flattenedMap.put(splitKey.trim(), value.trim());
+ }
+ }
+ }
+ return flattenedMap;
+ }
+
+ private static MapPartitionsFunction<Row, FileMetadata> toFileMetadata(
+ Map<String, String> equalSchemesMap, Map<String, String> equalAuthoritiesMap) {
+ return rows ->
+ Iterators.transform(
Review Comment:
Same.
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
Nit: could we quickly fix the formatting? It should be on same line like in Spark 3.2?
`Iterator<String> transformed = Iterators.transform(rows, row -> {`
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
+ row -> {
+ FileMetadata actual = row._1;
+ FileMetadata valid = row._2;
+
+ if (valid == null) {
+ return actual.location;
+ }
+
+ boolean schemeMatch =
+ Strings.isNullOrEmpty(valid.scheme)
+ || valid.scheme.equalsIgnoreCase(actual.scheme);
+ boolean authorityMatch =
+ Strings.isNullOrEmpty(valid.authority)
+ || valid.authority.equalsIgnoreCase(actual.authority);
+
+ if ((!schemeMatch || !authorityMatch) && mode == PrefixMismatchMode.DELETE) {
+ return actual.location;
+ } else {
+ if (!schemeMatch) {
+ conflicts.add(Pair.of(valid.scheme, actual.scheme));
+ }
+ if (!authorityMatch) {
+ conflicts.add(Pair.of(valid.authority, actual.authority));
+ }
+ }
+
+ return null;
+ });
+ return Iterators.filter(transformed, Objects::nonNull);
+ };
+ }
+
+ private static Map<String, String> flattenMap(Map<String, String> map) {
Review Comment:
Should this method be in the same place, after the first findOrphanFiles(), for easier cherry-pick of further changes?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1569323454
Can anyone take a look at this PR?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1266249371
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
+ row -> {
+ FileMetadata actual = row._1;
+ FileMetadata valid = row._2;
+
+ if (valid == null) {
+ return actual.location;
+ }
+
+ boolean schemeMatch =
+ Strings.isNullOrEmpty(valid.scheme)
+ || valid.scheme.equalsIgnoreCase(actual.scheme);
+ boolean authorityMatch =
+ Strings.isNullOrEmpty(valid.authority)
+ || valid.authority.equalsIgnoreCase(actual.authority);
+
+ if ((!schemeMatch || !authorityMatch) && mode == PrefixMismatchMode.DELETE) {
+ return actual.location;
+ } else {
+ if (!schemeMatch) {
+ conflicts.add(Pair.of(valid.scheme, actual.scheme));
+ }
+ if (!authorityMatch) {
+ conflicts.add(Pair.of(valid.authority, actual.authority));
+ }
+ }
+
+ return null;
+ });
+ return Iterators.filter(transformed, Objects::nonNull);
+ };
+ }
+
+ private static Map<String, String> flattenMap(Map<String, String> map) {
+ Map<String, String> flattenedMap = Maps.newHashMap();
+ if (map != null) {
+ for (String key : map.keySet()) {
+ String value = map.get(key);
+ for (String splitKey : COMMA.split(key)) {
+ flattenedMap.put(splitKey.trim(), value.trim());
+ }
+ }
+ }
+ return flattenedMap;
+ }
+
+ private static MapPartitionsFunction<Row, FileMetadata> toFileMetadata(
+ Map<String, String> equalSchemesMap, Map<String, String> equalAuthoritiesMap) {
+ return rows ->
+ Iterators.transform(
Review Comment:
Same, can be one line like Spark 3.2
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1266380690
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
```
> Task :iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck FAILED
FAILURE: Build failed with an exception.
* What went wrong:
Execution failed for task ':iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck'.
> The following files had format violations:
src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java
@@ -394,7 +394,9 @@
······PrefixMismatchMode·mode,·SetAccumulator<Pair<String,·String>>·conflicts)·{
····return·rows·->·{
······Iterator<String>·transformed·=
-··········Iterators.transform(rows,·row·->·{
+··········Iterators.transform(
+··············rows,
+··············row·->·{
················FileMetadata·actual·=·row._1;
················FileMetadata·valid·=·row._2;
@@ -429,7 +431,9 @@
··private·static·MapPartitionsFunction<Row,·FileMetadata>·toFileMetadata(
······Map<String,·String>·equalSchemesMap,·Map<String,·String>·equalAuthoritiesMap)·{
····return·rows·->
-········Iterators.transform(rows,·row·->·{
+········Iterators.transform(
+············rows,
+············row·->·{
··············String·location·=·row.getString(0);
··············URI·uri·=·new·Path(location).toUri();
··············String·scheme·=·equalSchemesMap.getOrDefault(uri.getScheme(),·uri.getScheme());
Run './gradlew :iceberg-spark:iceberg-spark-3.1_2.12:spotlessApply' to fix these violations.
```
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
```
> Task :iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck FAILED
FAILURE: Build failed with an exception.
* What went wrong:
Execution failed for task ':iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck'.
> The following files had format violations:
src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java
@@ -394,7 +394,9 @@
······PrefixMismatchMode·mode,·SetAccumulator<Pair<String,·String>>·conflicts)·{
····return·rows·->·{
······Iterator<String>·transformed·=
-··········Iterators.transform(rows,·row·->·{
+··········Iterators.transform(
+··············rows,
+··············row·->·{
················FileMetadata·actual·=·row._1;
················FileMetadata·valid·=·row._2;
@@ -429,7 +431,9 @@
··private·static·MapPartitionsFunction<Row,·FileMetadata>·toFileMetadata(
······Map<String,·String>·equalSchemesMap,·Map<String,·String>·equalAuthoritiesMap)·{
····return·rows·->
-········Iterators.transform(rows,·row·->·{
+········Iterators.transform(
+············rows,
+············row·->·{
··············String·location·=·row.getString(0);
··············URI·uri·=·new·Path(location).toUri();
··············String·scheme·=·equalSchemesMap.getOrDefault(uri.getScheme(),·uri.getScheme());
Run './gradlew :iceberg-spark:iceberg-spark-3.1_2.12:spotlessApply' to fix these violations.
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1267295678
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
Hm, I dont how did it work in : https://github.com/apache/iceberg/pull/4652/files, was that before spotless? its ok then.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1640979221
Thanks @manuzhang
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1266251513
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
+ row -> {
+ FileMetadata actual = row._1;
+ FileMetadata valid = row._2;
+
+ if (valid == null) {
+ return actual.location;
+ }
+
+ boolean schemeMatch =
+ Strings.isNullOrEmpty(valid.scheme)
+ || valid.scheme.equalsIgnoreCase(actual.scheme);
+ boolean authorityMatch =
+ Strings.isNullOrEmpty(valid.authority)
+ || valid.authority.equalsIgnoreCase(actual.authority);
+
+ if ((!schemeMatch || !authorityMatch) && mode == PrefixMismatchMode.DELETE) {
+ return actual.location;
+ } else {
+ if (!schemeMatch) {
+ conflicts.add(Pair.of(valid.scheme, actual.scheme));
+ }
+ if (!authorityMatch) {
+ conflicts.add(Pair.of(valid.authority, actual.authority));
+ }
+ }
+
+ return null;
+ });
+ return Iterators.filter(transformed, Objects::nonNull);
+ };
+ }
+
+ private static Map<String, String> flattenMap(Map<String, String> map) {
Review Comment:
Should this method be in the same place, after the first findOrphanFiles(), as Spark 3.2, for easier cherry-pick of further changes?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1641393552
@szehon-ho I'm wondering we still need to backport #5495 since from [Anton's comment](https://github.com/apache/iceberg/pull/5495#issuecomment-1219689082) there's no big difference after the change.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1266380690
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
The previous formatting no longer passes spotless check.
```
> Task :iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck FAILED
FAILURE: Build failed with an exception.
* What went wrong:
Execution failed for task ':iceberg-spark:iceberg-spark-3.1_2.12:spotlessJavaCheck'.
> The following files had format violations:
src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java
@@ -394,7 +394,9 @@
······PrefixMismatchMode·mode,·SetAccumulator<Pair<String,·String>>·conflicts)·{
····return·rows·->·{
······Iterator<String>·transformed·=
-··········Iterators.transform(rows,·row·->·{
+··········Iterators.transform(
+··············rows,
+··············row·->·{
················FileMetadata·actual·=·row._1;
················FileMetadata·valid·=·row._2;
@@ -429,7 +431,9 @@
··private·static·MapPartitionsFunction<Row,·FileMetadata>·toFileMetadata(
······Map<String,·String>·equalSchemesMap,·Map<String,·String>·equalAuthoritiesMap)·{
····return·rows·->
-········Iterators.transform(rows,·row·->·{
+········Iterators.transform(
+············rows,
+············row·->·{
··············String·location·=·row.getString(0);
··············URI·uri·=·new·Path(location).toUri();
··············String·scheme·=·equalSchemesMap.getOrDefault(uri.getScheme(),·uri.getScheme());
Run './gradlew :iceberg-spark:iceberg-spark-3.1_2.12:spotlessApply' to fix these violations.
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1554110654
@aokolnychyi @karuppayya Please help review the back-port. I think it's critical for Spark 3.1 users as the prefix mismatched has caused failure of maintenance jobs (e.g. ExpireSnapshots) in our cluster.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1639454331
Any more comments here?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] manuzhang commented on pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "manuzhang (via GitHub)" <gi...@apache.org>.
manuzhang commented on PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#issuecomment-1639738170
@szehon-ho thanks for review. I'll backport #5495 in a follow-up PR.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
[GitHub] [iceberg] szehon-ho commented on a diff in pull request #7653: Spark 3.1: Add prefix mismatch mode for deleting orphan files
Posted by "szehon-ho (via GitHub)" <gi...@apache.org>.
szehon-ho commented on code in PR #7653:
URL: https://github.com/apache/iceberg/pull/7653#discussion_r1267295678
##########
spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/actions/BaseDeleteOrphanFilesSparkAction.java:
##########
@@ -296,4 +334,160 @@ private static FlatMapFunction<Iterator<String>, String> listDirsRecursively(
return files.iterator();
};
}
+
+ @VisibleForTesting
+ static List<String> findOrphanFiles(
+ SparkSession spark,
+ Dataset<Row> actualFileDF,
+ Dataset<Row> validFileDF,
+ Map<String, String> equalSchemes,
+ Map<String, String> equalAuthorities,
+ PrefixMismatchMode prefixMismatchMode) {
+ Dataset<FileMetadata> actualFileMetadataDS =
+ actualFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+ Dataset<FileMetadata> validFileMetadataDS =
+ validFileDF.mapPartitions(
+ toFileMetadata(equalSchemes, equalAuthorities), Encoders.bean(FileMetadata.class));
+
+ SetAccumulator<Pair<String, String>> conflicts = new SetAccumulator<>();
+ spark.sparkContext().register(conflicts);
+
+ Column joinCond = actualFileMetadataDS.col("path").equalTo(validFileMetadataDS.col("path"));
+
+ List<String> orphanFiles =
+ actualFileMetadataDS
+ .joinWith(validFileMetadataDS, joinCond, "leftouter")
+ .mapPartitions(findOrphanFiles(prefixMismatchMode, conflicts), Encoders.STRING())
+ .collectAsList();
+
+ if (prefixMismatchMode == PrefixMismatchMode.ERROR && !conflicts.value().isEmpty()) {
+ throw new ValidationException(
+ "Unable to determine whether certain files are orphan. "
+ + "Metadata references files that match listed/provided files except for authority/scheme. "
+ + "Please, inspect the conflicting authorities/schemes and provide which of them are equal "
+ + "by further configuring the action via equalSchemes() and equalAuthorities() methods. "
+ + "Set the prefix mismatch mode to 'NONE' to ignore remaining locations with conflicting "
+ + "authorities/schemes or to 'DELETE' iff you are ABSOLUTELY confident that remaining conflicting "
+ + "authorities/schemes are different. It will be impossible to recover deleted files. "
+ + "Conflicting authorities/schemes: %s.",
+ conflicts.value());
+ }
+
+ return orphanFiles;
+ }
+
+ private static MapPartitionsFunction<Tuple2<FileMetadata, FileMetadata>, String> findOrphanFiles(
+ PrefixMismatchMode mode, SetAccumulator<Pair<String, String>> conflicts) {
+ return rows -> {
+ Iterator<String> transformed =
+ Iterators.transform(
+ rows,
Review Comment:
Hm, I dont how did it work in : https://github.com/apache/iceberg/pull/4652/files its ok then.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org