[
https://issues.apache.org/jira/browse/IGNITE-7655?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16376422#comment-16376422
]
Nikolay Izhikov edited comment on IGNITE-7655 at 2/26/18 6:01 AM:
------------------------------------------------------------------
[~dmagda]
> What does appending mean here? As far as I understand, you're just going to
> insert new data ignoring the data which keys are already in the table.
Appends mean we will not execute {{DROP TABLE}}/{{CREATE TABLE}}.
Since all inserts are executed through IgniteDataStreamer behaviour will be
like this:
* {{OPTION_STREAMER_ALLOW_OVERWRITE=false}} - existing keys will be ignored.
* {{OPTION_STREAMER_ALLOW_OVERWRITE=true}} - existing key/value will
overwritten with DataFrame data.
> which mode is a default?
ErrorIfExsists
> Is there a mode that updates existing rows, inserts new from the data frame
> and does NOT delete the rows that were in the table initially?
{{SaveMode.Append}} + {{OPTION_STREAMER_ALLOW_OVERWRITE=true}}
> How do I set a mode? You start talking about the data frames right away but
> do not explain how to use them.
One can use {{dataFrame.write.mode(SaveMode.Append)}}. It's a part of Spark
API.
OK, I will add an example to the page.
> I would start "Saving Data Frame" section with the explanation of how all
> this works
OK. Will do.
was (Author: nizhikov):
[~dmagda]
> What does appending mean here? As far as I understand, you're just going to
> insert new data ignoring the data which keys are already in the table.
Appends mean we will not execute `DROP TABLE`/`CREATE TABLE`.
Since all inserts are executed through IgniteDataStreamer behaviour will be
like this:
* `OPTION_STREAMER_ALLOW_OVERWRITE=false` - existing keys will be ignored.
* `OPTION_STREAMER_ALLOW_OVERWRITE=true` - existing key/value will overwritten
with DataFrame data.
> which mode is a default?
ErrorIfExsists
> Is there a mode that updates existing rows, inserts new from the data frame
> and does NOT delete the rows that were in the table initially?
`SaveMode.Append` + `OPTION_STREAMER_ALLOW_OVERWRITE=true`
> How do I set a mode? You start talking about the data frames right away but
> do not explain how to use them.
One can use `dataFrame.write.mode(SaveMode.Append)`. It's a part of Spark API.
OK, I will add an example to the page.
> I would start "Saving Data Frame" section with the explanation of how all
> this works
OK. Will do.
> Spark Data Frames: saving data frames into Ignite needs to be documented
> ------------------------------------------------------------------------
>
> Key: IGNITE-7655
> URL: https://issues.apache.org/jira/browse/IGNITE-7655
> Project: Ignite
> Issue Type: Bug
> Components: documentation, spark
> Reporter: Nikolay Izhikov
> Assignee: Nikolay Izhikov
> Priority: Major
> Fix For: 2.4
>
>
> Once IGNITE-7337 is ready for merge.
> This new feature of Ignite needs to be documented.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
