deniskuzZ commented on code in PR #4431:
URL: https://github.com/apache/hive/pull/4431#discussion_r1238450232
##########
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergStorageHandler.java:
##########
@@ -493,20 +501,47 @@ private String getStatsSource() {
}
private Path getStatsPath(Table table) {
- return new Path(table.location() + STATS + table.name() +
table.currentSnapshot().snapshotId());
+ return getStatsPath(table, table.currentSnapshot().snapshotId());
+ }
+
+ private Path getStatsPath(Table table, long snapshotId) {
+ return new Path(table.location() + STATS + table.name() + snapshotId);
}
- private void invalidateStats(Path statsPath) {
+ private void checkAndMergeOrInvalidateStats(ColumnStatistics statsObjNew,
+ org.apache.hadoop.hive.ql.metadata.Table hmsTable) {
+ Table tbl = IcebergTableUtil.getTable(conf, hmsTable.getTTable());
+ Path statsPath = getStatsPath(tbl);
try {
FileSystem fs = statsPath.getFileSystem(conf);
if (fs.exists(statsPath)) {
+ // Analyze table and stats updater thread
fs.delete(statsPath, true);
+ } else {
+ // Insert queries
+ mergeStats(statsObjNew, tbl, hmsTable);
}
} catch (IOException e) {
LOG.error("Failed to invalidate stale column stats: {}", e.getMessage());
}
}
+ private void mergeStats(ColumnStatistics statsObjNew, Table tbl,
org.apache.hadoop.hive.ql.metadata.Table hmsTable) {
+ if (tbl.currentSnapshot().parentId() != null) {
+ long previousSnapshotId = tbl.currentSnapshot().parentId();
+ if (canProvideColStatistics(hmsTable, previousSnapshotId)) {
+ ColumnStatistics statsObjOld = readColStats(tbl, getStatsPath(tbl,
previousSnapshotId).toString());
+ if (statsObjOld != null) {
+ try {
+ MetaStoreServerUtils.mergeColStats(statsObjNew, statsObjOld);
+ } catch (InvalidObjectException e) {
+ LOG.error("Unable to merge column stats: ", e.getMessage());
Review Comment:
would we mark stats as accurate in case of an exception?
--
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]
