ajantha-bhat commented on code in PR #6091:
URL: https://github.com/apache/iceberg/pull/6091#discussion_r1123027978
##########
spark/v3.3/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestExpireSnapshotsProcedure.java:
##########
@@ -384,10 +402,105 @@ public void
testExpireSnapshotsProcedureWorksWithSqlComments() {
+ " table => '%s',"
+ " retain_last => 1)";
List<Object[]> output = sql(callStatement, catalogName, currentTimestamp,
tableIdent);
- assertEquals("Procedure output must match", ImmutableList.of(row(0L, 0L,
0L, 0L, 1L)), output);
+ assertEquals(
+ "Procedure output must match", ImmutableList.of(row(0L, 0L, 0L, 0L,
1L, 0L)), output);
table.refresh();
Assert.assertEquals("Should be 1 snapshot remaining", 1,
Iterables.size(table.snapshots()));
}
+
+ @Test
+ public void testExpireSnapshotsWithStatisticFiles() throws Exception {
+ sql(
+ "CREATE TABLE %s USING iceberg "
+ + "TBLPROPERTIES('format-version'='2') "
+ + "AS SELECT 10 int, 'abc' data",
+ tableName);
+ Table table = Spark3Util.loadIcebergTable(spark, tableName);
+ String statsFileLocation1 = statsFileLocation(table.location());
+ StatisticsFile statisticsFile1 =
+ writeStatsFile(
+ table.currentSnapshot().snapshotId(),
+ table.currentSnapshot().sequenceNumber(),
+ statsFileLocation1,
+ table.io());
+ table.updateStatistics().setStatistics(statisticsFile1.snapshotId(),
statisticsFile1).commit();
+
+ sql("INSERT INTO %s SELECT 20, 'def'", tableName);
+ table = Spark3Util.loadIcebergTable(spark, tableName);
+ String statsFileLocation2 = statsFileLocation(table.location());
+ StatisticsFile statisticsFile2 =
+ writeStatsFile(
+ table.currentSnapshot().snapshotId(),
+ table.currentSnapshot().sequenceNumber(),
+ statsFileLocation2,
+ table.io());
+ table.updateStatistics().setStatistics(statisticsFile2.snapshotId(),
statisticsFile2).commit();
+
+ Timestamp currentTimestamp =
Timestamp.from(Instant.ofEpochMilli(System.currentTimeMillis()));
+ List<Object[]> output =
+ sql(
+ "CALL %s.system.expire_snapshots("
+ + "older_than => TIMESTAMP '%s',"
+ + "table => '%s',"
+ + "retain_last => 1, "
+ + "stream_results => true)",
+ catalogName, currentTimestamp, tableIdent);
+ Assertions.assertThat(output.get(0)[5]).as("should be 1 deleted statistics
file").isEqualTo(1L);
+
+ table = Spark3Util.loadIcebergTable(spark, tableName);
+ List<StatisticsFile> statsWithSnapshotId1 =
+ table.statisticsFiles().stream()
+ .filter(statisticsFile -> statisticsFile.snapshotId() ==
statisticsFile1.snapshotId())
+ .collect(Collectors.toList());
+ Assertions.assertThat(statsWithSnapshotId1)
+ .as(
+ "Statistics file entry in TableMetadata should be deleted for the
snapshot %s",
+ statisticsFile1.snapshotId())
+ .isEmpty();
+ Assertions.assertThat(table.statisticsFiles())
+ .as(
+ "Statistics file entry in TableMetadata should be present for the
snapshot %s",
+ statisticsFile2.snapshotId())
+ .extracting(StatisticsFile::snapshotId)
+ .containsExactly(statisticsFile2.snapshotId());
+
+ Assertions.assertThat(new File(statsFileLocation1))
+ .as("Statistics file should not exist for snapshot %s",
statisticsFile1.snapshotId())
+ .doesNotExist();
+
+ Assertions.assertThat(new File(statsFileLocation2))
+ .as("Statistics file should exist for snapshot %s",
statisticsFile2.snapshotId())
+ .exists();
+ }
+
+ private StatisticsFile writeStatsFile(
+ long snapshotId, long snapshotSequenceNumber, String statsLocation,
FileIO fileIO)
+ throws IOException {
+ try (PuffinWriter puffinWriter =
Puffin.write(fileIO.newOutputFile(statsLocation)).build()) {
+ puffinWriter.add(
+ new Blob(
+ "some-blob-type",
+ ImmutableList.of(1),
+ snapshotId,
+ snapshotSequenceNumber,
+ ByteBuffer.wrap("blob
content".getBytes(StandardCharsets.UTF_8))));
+ puffinWriter.finish();
+
+ return new GenericStatisticsFile(
+ snapshotId,
+ statsLocation,
+ puffinWriter.fileSize(),
+ puffinWriter.footerSize(),
+ puffinWriter.writtenBlobsMetadata().stream()
+ .map(GenericBlobMetadata::from)
+ .collect(ImmutableList.toImmutableList()));
+ }
+ }
+
+ private String statsFileLocation(String tableLocation) {
Review Comment:
I have got a comment to not use `File` and `Uri` in the core impl PR for
this logic.
https://github.com/apache/iceberg/pull/6090#discussion_r1053490603
Hence, I handled it like this. In the core test case `table.location()`
returns a path without `file:`. But this module returns with `file:`. Many
other places in this module we are replacing ` file:` in table location.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
