[
https://issues.apache.org/jira/browse/DRILL-5033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17653150#comment-17653150
]
ASF GitHub Bot commented on DRILL-5033:
---------------------------------------
cgivre commented on PR #2731:
URL: https://github.com/apache/drill/pull/2731#issuecomment-1367977377
> Thanks @cgivre for the clarification, but suppose the assumption that
considering nulls as strings would solve the issue, were the changes i made
(over the class JSONReader.java) adequate (should the methods be changed as i
did)? i see that some tests didn't pass.
For us to merge a pull request, all the unit tests have to pass. (Or be
modified with an explanation of why they are being modified) Drill is a very
complex beast with a lot of dependencies so even small changes can break things
you didn't intend to. Believe me... I know from experience ;-)
One other thing to note is that there is another option
`drill.exec.functions.cast_empty_string_to_null`. Setting this to true forces
empty strings to be treated as `null`. This can have some unintended side
effects, but might also help you out.
> Query on JSON that has null as value for each key
> -------------------------------------------------
>
> Key: DRILL-5033
> URL: https://issues.apache.org/jira/browse/DRILL-5033
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - JSON
> Affects Versions: 1.9.0
> Reporter: Khurram Faraaz
> Priority: Major
>
> Drill 1.9.0 git commit ID : 83513daf
> Drill returns same result with or without `store.json.all_text_mode`=true
> Note that each key in the JSON has null as its value.
> [root@cent01 null_eq_joins]# cat right_all_nulls.json
> {
> "intKey" : null,
> "bgintKey": null,
> "strKey": null,
> "boolKey": null,
> "fltKey": null,
> "dblKey": null,
> "timKey": null,
> "dtKey": null,
> "tmstmpKey": null,
> "intrvldyKey": null,
> "intrvlyrKey": null
> }
> [root@cent01 null_eq_joins]#
> Querying the above JSON file results in null as query result.
> - We should see each of the keys in the JSON as a column in query result.
> - And in each column the value should be a null value.
> Current behavior does not look right.
> {noformat}
> 0: jdbc:drill:schema=dfs.tmp> select * from `right_all_nulls.json`;
> +-------+
> | * |
> +-------+
> | null |
> +-------+
> 1 row selected (0.313 seconds)
> {noformat}
> Adding comment from [~julianhyde]
> IMHO it is similar but not the same as DRILL-1256. Worth logging an issue and
> let [~jnadeau] (or someone) put on the record what should be the behavior of
> an empty record (empty JSON map) when it is top-level (as in this case) or in
> a collection.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
