[
https://issues.apache.org/jira/browse/SPARK-32439?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Terry Kim updated SPARK-32439:
------------------------------
Description:
We need a mechanism to override the datasource implementation via configuration.
For example, suppose I have a custom CSV datasource implementation called
"my_csv". One way to use it is:
{code}
val df = spark.read.format("my_csv").load(...)
{code}
Since the source data is the same format (CSV), you should be able to override
the default implementation.
One proposal is to do the following:
{code}
spark.conf.set("spark.sql.datasource.override.csv", "my_csv") val df =
spark.read.csv(...)
{code}
This has a benefit that the user doesn't have to change the application code to
try out a new datasource implementation for the same source format.
was:
We need a mechanism to override the datasource implementation via configuration.
For example, suppose I have a custom CSV datasource implementation called
"my_csv". One way to use it is:
val df = spark.read.format("my_csv").load(...)
Since the source data is the same format (CSV), you should be able to override
the default implementation.
One proposal is to do the following:
spark.conf.set("spark.sql.datasource.override.csv", "my_csv") val df =
spark.read.csv(...)
This has a benefit that the user doesn't have to change the application code to
try out a new datasource implementation for the same source format.
> Override datasource implementation during look up via configuration
> -------------------------------------------------------------------
>
> Key: SPARK-32439
> URL: https://issues.apache.org/jira/browse/SPARK-32439
> Project: Spark
> Issue Type: New Feature
> Components: SQL
> Affects Versions: 3.1.0
> Reporter: Terry Kim
> Priority: Minor
>
> We need a mechanism to override the datasource implementation via
> configuration.
> For example, suppose I have a custom CSV datasource implementation called
> "my_csv". One way to use it is:
> {code}
> val df = spark.read.format("my_csv").load(...)
> {code}
> Since the source data is the same format (CSV), you should be able to
> override the default implementation.
> One proposal is to do the following:
> {code}
> spark.conf.set("spark.sql.datasource.override.csv", "my_csv") val df =
> spark.read.csv(...)
> {code}
> This has a benefit that the user doesn't have to change the application code
> to try out a new datasource implementation for the same source format.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
