[
https://issues.apache.org/jira/browse/SPARK-41741?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Apache Spark reassigned SPARK-41741:
------------------------------------
Assignee: (was: Apache Spark)
> [SQL] ParquetFilters StringStartsWith push down matching string do not use
> UTF-8
> --------------------------------------------------------------------------------
>
> Key: SPARK-41741
> URL: https://issues.apache.org/jira/browse/SPARK-41741
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 3.4.0
> Reporter: Jiale He
> Priority: Major
> Attachments: image-2022-12-28-18-00-00-861.png,
> image-2022-12-28-18-00-21-586.png, image-2023-01-09-11-10-31-262.png,
> image-2023-01-09-18-27-53-479.png,
> part-00000-30432312-7cdb-43ef-befe-93bcfd174878-c000.snappy.parquet
>
>
> Hello ~
>
> I found a problem, but there are two ways to solve it.
>
> The parquet filter is pushed down. When using the like '***%' statement to
> query, if the system default encoding is not UTF-8, it may cause an error.
>
> There are two ways to bypass this problem as far as I know
> 1. spark.executor.extraJavaOptions="-Dfile.encoding=UTF-8"
> 2. spark.sql.parquet.filterPushdown.string.startsWith=false
>
> The following is the information to reproduce this problem
> The parquet sample file is in the attachment
> {code:java}
> spark.read.parquet("file:///home/kylin/hjldir/part-00000-30432312-7cdb-43ef-befe-93bcfd174878-c000.snappy.parquet").createTempView("tmp”)
> spark.sql("select * from tmp where `1` like '啦啦乐乐%'").show(false) {code}
> !image-2022-12-28-18-00-00-861.png|width=879,height=430!
>
> !image-2022-12-28-18-00-21-586.png|width=799,height=731!
>
> I think the correct code should be:
> {code:java}
> private val strToBinary =
> Binary.fromReusedByteArray(v.getBytes(StandardCharsets.UTF_8)) {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
