[ 
https://issues.apache.org/jira/browse/HIVE-26401?focusedWorklogId=796587&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-796587
 ]

ASF GitHub Bot logged work on HIVE-26401:
-----------------------------------------

                Author: ASF GitHub Bot
            Created on: 30/Jul/22 09:08
            Start Date: 30/Jul/22 09:08
    Worklog Time Spent: 10m 
      Work Description: wecharyu commented on PR #3447:
URL: https://github.com/apache/hive/pull/3447#issuecomment-1200121540

   > Better not to mix two thing: refactor & log stuff together. Can you just 
restrict this PR to just Log refinement
   
   Hi @ayushtkn , do you mean do not refactor `validatePartition()` method in 
this PR? I make this refactor to get partition keys from table, how about 
logging partition information in the format `tableName[part_val1,part_val2...]` 
without refactor?




Issue Time Tracking
-------------------

    Worklog Id:     (was: 796587)
    Time Spent: 20m  (was: 10m)

> Refine the log of add_partitions if the partition already exists
> ----------------------------------------------------------------
>
>                 Key: HIVE-26401
>                 URL: https://issues.apache.org/jira/browse/HIVE-26401
>             Project: Hive
>          Issue Type: Improvement
>          Components: Hive
>    Affects Versions: 4.0.0-alpha-1
>            Reporter: Wechar
>            Assignee: Wechar
>            Priority: Minor
>              Labels: pull-request-available
>             Fix For: 4.0.0-alpha-2
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Currently {{*add_partitions_xxx*}} will log the complete information of a 
> partition if it already exists, see in 
> [HMSHandler.java#L4320|https://github.com/apache/hive/blob/e3751ab545370f9b252d0b4a07bc315037541a95/standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/HMSHandler.java#L4320]:
> {code:java}
>     if (!shouldAdd) {
>       LOG.info("Not adding partition {} as it already exists", part);
>       return false;
>     }
> {code}
> It will print a long message including the columns of this partition, we 
> think it is unnecessary based on the following two points:
> {color:red}1. The long message is redundant.{color}
> We can get enough information from just 
> *cat_name.db_name.tbl_name[part_col1=part_val1/part_col2=part_val2...]*
> {color:red}2. The long message is not friendly to save and query.{color}
> This log message will take up a large log space especially when the user need 
> to execute *MSCK REPAIR TABLE* operation regularly because the old partition 
> must be already existed. 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to