Todd Lipcon has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/13334 )

Change subject: acid: Filter unwanted files based on ACID state.
......................................................................


Patch Set 14:

(4 comments)

I think Sudhanshu's AFK this afternoon so will try to rev the patch in his 
absence.

http://gerrit.cloudera.org:8080/#/c/13334/14/fe/src/main/java/org/apache/impala/util/AcidUtils.java
File fe/src/main/java/org/apache/impala/util/AcidUtils.java:

http://gerrit.cloudera.org:8080/#/c/13334/14/fe/src/main/java/org/apache/impala/util/AcidUtils.java@48
PS14, Line 48: (?:/.*)?
> I have deltas like this produced by compaction:
Good point. I'll see if I can add a test case for this one.


http://gerrit.cloudera.org:8080/#/c/13334/14/fe/src/main/java/org/apache/impala/util/AcidUtils.java@54
PS14, Line 54:        // Optional path suffix.
             :        "(?:/.*)?");
> Is this optional? We should only match files inside directories as far as I
yea, I guess in contrast to base patterns, the existence of an empty delta 
directory doesn't actually matter, so we could skip over the directories and 
only match on the files. That said, I don't think there's harm in matching the 
directories and it seems less surprising if this regex matches the dir (same as 
above)


http://gerrit.cloudera.org:8080/#/c/13334/14/fe/src/main/java/org/apache/impala/util/AcidUtils.java@105
PS14, Line 105:     public boolean test(String dirPath) {
> To support upgraded tables this should also accept any path without "/", so
Good point. Mind if I add a TODO to support upgraded tables? I'm not sure how 
to test this at the moment and this patch is already large.


http://gerrit.cloudera.org:8080/#/c/13334/14/fe/src/main/java/org/apache/impala/util/AcidUtils.java@182
PS14, Line 182:     // If the table was not transaction table then return all 
fds.
              :     if (validDescriptors.isEmpty()) {
              :       return fileDescriptors;
              :     }
> Can't we enter this if the table was truncated, but the old deltas are stil
Yea, I think that's true. Unfortunately right now due to 
https://issues.apache.org/jira/browse/HIVE-20137 and 
https://issues.apache.org/jira/browse/HIVE-21750 there doesn't seem to be a way 
to test this case at the moment. I'll fix it anyway and add a unit test.



--
To view, visit http://gerrit.cloudera.org:8080/13334
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Icf0aeb36e10c827ead59ed7f67e731199394fe8e
Gerrit-Change-Number: 13334
Gerrit-PatchSet: 14
Gerrit-Owner: Sudhanshu Arora <[email protected]>
Gerrit-Reviewer: Csaba Ringhofer <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Sudhanshu Arora <[email protected]>
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-Reviewer: Vihang Karajgaonkar <[email protected]>
Gerrit-Reviewer: Yongzhi Chen <[email protected]>
Gerrit-Reviewer: Zoltan Borok-Nagy <[email protected]>
Gerrit-Comment-Date: Wed, 22 May 2019 20:34:46 +0000
Gerrit-HasComments: Yes

Reply via email to