[
https://issues.apache.org/jira/browse/SPARK-19257?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15832935#comment-15832935
]
Song Jun edited comment on SPARK-19257 at 1/22/17 11:22 AM:
------------------------------------------------------------
I found it that `CatalogTablePartition` and `CatalogTable` use the same class
`CatalogTableStorageFormat`.
if we change CatalogTableStorageFormat 's locationUri from `String` to `URI`,
this will failed in `CatalogTablePartition` because table partition location
can have whitespace, such as `2014-01-01 00%3A00%3A00`， while this is not a
legal URI (no encoded whitespace)
a)should we seperate the CatalogTableStorageFormat from CatalogTablePartition
and CatalogTable?
b)or we just check if the ` locationUri: String` is illegal when make a DDL
command, such as ` alter table set location xx` or `create table t options
(path xx) `
Hive's implement is like b) :
https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/parse/DDLSemanticAnalyzer.java#L1553
https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java#L3732
HIVE-0.12 the table's location method's paramenter Type is URI, after that
version change to Path
https://github.com/apache/hive/blob/branch-0.12/ql/src/java/org/apache/hadoop/hive/ql/metadata/Table.java#L501
this is the same issue in hive:
https://issues.apache.org/jira/browse/HIVE-6185
[~cloud_fan] [~smilegator]
was (Author: windpiger):
I found it that `CatalogTablePartition` and `CatalogTable` use the same class
`CatalogTableStorageFormat`.
if we change CatalogTableStorageFormat 's locationUri from `String` to `URI`,
this will failed in `CatalogTablePartition` because table partition location
can have whitespace, such as `2014-01-01 00%3A00%3A00`， while this is not a
legal URI (no encoded whitespace)
a)should we seperate the CatalogTableStorageFormat from CatalogTablePartition
and CatalogTable?
b)or we just check if the ` locationUri: String` is illegal when make a DDL
command, such as ` alter table set location xx` or `create table t options
(path xx) `
Hive's implement is like b) :
https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/parse/DDLSemanticAnalyzer.java#L1553
https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/exec/DDLTask.java#L3732
[~cloud_fan] [~smilegator]
> The type of CatalogStorageFormat.locationUri should be java.net.URI instead
> of String
> -------------------------------------------------------------------------------------
>
> Key: SPARK-19257
> URL: https://issues.apache.org/jira/browse/SPARK-19257
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Reporter: Wenchen Fan
>
> Currently we treat `CatalogStorageFormat.locationUri` as URI string and
> always convert it to path by `new Path(new URI(locationUri))`
> It will be safer if we can make the type of
> `CatalogStorageFormat.locationUri` java.net.URI. We should finish the TODO in
> https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/catalog/interface.scala#L50-L52
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
