[
https://issues.apache.org/jira/browse/SPARK-17749?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15538938#comment-15538938
]
Dongjoon Hyun commented on SPARK-17749:
---------------------------------------
Hi, [~andreasdamm].
Thank you for reporting issue. I saw that 2.0.0 raises the following as you
mentioned.
```
org.apache.spark.sql.AnalysisException: cannot resolve
'`sf_opportunity.currencyisocode`' given input columns: [isocode, id]; line 1
pos
```
For current master branch (Spark 2.1.0-SNAPSHOT), it's a little bit different.
```
org.apache.spark.sql.catalyst.parser.ParseException:
extraneous input 'ON' expecting
```
For Spark 2.0.1-rc4, I didn't test on that, but I guess it has the similar
errors.
Could you try with the following query by replacing the second 'ON' with 'AND'?
The following query is tested on Spark 2.0.0 and 2.1.0-SNAPSHOT.
```
SELECT 0
FROM `sf_datedconversionrate2` AS `sf_datedconversionrate`
LEFT JOIN `sf_account2` AS `sf_account`
LEFT JOIN `sf_opportunity2` AS `sf_opportunity`
ON `sf_account`.`id` = `sf_opportunity`.`accountid` AND
`sf_datedconversionrate`.`isocode` = `sf_opportunity`.`currencyisocode`
```
> Unresolved columns when nesting SQL join clauses
> ------------------------------------------------
>
> Key: SPARK-17749
> URL: https://issues.apache.org/jira/browse/SPARK-17749
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.0.0
> Reporter: Andreas Damm
>
> Given tables
> CREATE TABLE `sf_datedconversionrate2`(`isocode` string)
> CREATE TABLE `sf_opportunity2`(`currencyisocode` string, `accountid` string)
> CREATE TABLE `sf_account2`(`id` string)
> the following SQL will cause an analysis exception (cannot resolve
> '`sf_opportunity.currencyisocode`' given input columns: [isocode, id])
> SELECT 0
> FROM `sf_datedconversionrate2` AS `sf_datedconversionrate`
> LEFT JOIN `sf_account2` AS `sf_account`
> LEFT JOIN `sf_opportunity2` AS `sf_opportunity`
> ON `sf_account`.`id` = `sf_opportunity`.`accountid`
> ON `sf_datedconversionrate`.`isocode` =
> `sf_opportunity`.`currencyisocode`
> even though all columns referred to in the conditions should be in scope.
> Re-ordering the join and on clauses will make it work
> SELECT 0
> FROM `sf_datedconversionrate2` AS `sf_datedconversionrate`
> LEFT JOIN `sf_opportunity2` AS `sf_opportunity`
> LEFT JOIN `sf_account2` AS `sf_account`
> ON `sf_account`.`id` = `sf_opportunity`.`accountid`
> ON `sf_datedconversionrate`.`isocode` =
> `sf_opportunity`.`currencyisocode`
> but the original should work also.
--
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
