[ 
https://issues.apache.org/jira/browse/CASSANDRA-16818?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17390634#comment-17390634
 ] 

Andres de la Peña commented on CASSANDRA-16818:
-----------------------------------------------

There were some CI test failures while checking {{assertSchemaNotLoaded}} with 
Java 11 that neither [~e.dimitrova] nor me have been able to reproduce locally. 
It seems that the loading of {{Schema}} class in the test cases that use 
{{TableMetadata}} might or might not load that class depending on the JVM 
version. I have changed the tests back to their previous form where that check 
is only done for the test cases where {{TableMetadata}} is used, adding a brief 
comment warning about the issue:
||PR||CI||
|[4.0|https://github.com/apache/cassandra/pull/1121]|[j8|https://app.circleci.com/pipelines/github/adelapena/cassandra/721/workflows/4decc78f-a2e5-43c0-90ee-618ad3c0d678]
 
[j11|https://app.circleci.com/pipelines/github/adelapena/cassandra/721/workflows/c72b39f7-eb58-4d93-a150-714244239825]|
|[trunk|https://github.com/apache/cassandra/pull/1122]|[j8|https://app.circleci.com/pipelines/github/adelapena/cassandra/722/workflows/fe5e7407-3af0-4a08-b196-434652ea8789]
 
[j11|https://app.circleci.com/pipelines/github/adelapena/cassandra/722/workflows/08f88bed-2879-4018-b76b-5dcc45e85ea3]|

> Fix usage instructions about sstabledump -k and -x options
> ----------------------------------------------------------
>
>                 Key: CASSANDRA-16818
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-16818
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Tool/sstable
>            Reporter: Andres de la Peña
>            Assignee: Andres de la Peña
>            Priority: Normal
>             Fix For: 3.0.x, 3.11.x, 4.0.x, 4.x
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> The options {{-k}} and {{-x}} of {{sstabledump}} admit multiple arguments, so 
> users can include or exclude multiple partitions. The intended usage is, for 
> example:
> {code:java}
> $ sstablepartitions <sstable_path> -k 1
> $ sstablepartitions <sstable_path> -k 1 2 3
> {code}
> However, the following command will fail:
> {code:java}
> $ sstablepartitions -k 1 <sstable_path>
> You must supply exactly one sstable
> usage: sstabledump <sstable file path> <options>
> Dump contents of given SSTable to standard output in JSON format.
>  -d         CQL row per line internal representation
>  -e         enumerate partition keys only
>  -k <arg>   Partition key
>  -t         Print raw timestamps instead of iso8601 date strings
>  -x <arg>   Excluded partition key
> {code}
> This command fails because the sstable path is considered a fourth partition 
> key, and the mandatory argument for the sstable path is missed. While I think 
> this behaviour is correct, it can be a bit confusing for users, especially 
> when the information about usage describes both {{-k}} and {{-x}} as 
> single-argument.
> I think that at least we should fix the description of the options to 
> indicate that there could be multiple included/excluded keys, and probably 
> improve the message about the missing sstable path when those options are 
> used.
> Alternatively we could modify the options to have a single argument and allow 
> to repeat them, so we could accept things like:
> {code:java}
> $ sstablepartitions -k 1 <sstable_path>
> $ sstablepartitions -k 1 -k 2 -k 3 <sstable_path>
> $ sstablepartitions <sstable_path> -k 1 -k 2 -k 3
> {code}
> The main downside of the latter approach is that the change in the syntax of 
> the command might cause compatibility issues. Also we would need to upgrade 
> {{commons-cli}} to at least 1.2 due to CLI-137.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org

Reply via email to