[
https://issues.apache.org/jira/browse/ARROW-17184?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dragoș Moldovan-Grünfeld updated ARROW-17184:
---------------------------------------------
Description:
*Update:* We currently do not print a plan for nested queries as that would
involve the execution of the inner most query / queries. That is likely the
reason why, in the chunk below, we do not see the filter and arrange nodes.
==========================
*Original description:*
Now that we can print an ExecPlan (ARROW-15016), we can investigate unexpected
outputs. With the following chunk of code:
{code:r}
mtcars %>%
arrow_table() %>%
select(mpg, wt, cyl) %>%
filter(mpg > 20) %>%
arrange(desc(wt)) %>%
head(3) %>%
show_exec_plan()
#> ExecPlan with 3 nodes:
#> 2:SinkNode{}
#> 1:ProjectNode{projection=[mpg, wt, cyl]}
#> 0:SourceNode{}
{code}
* FilterNode disappears when head/tail are involved +
* we do not have additional information regarding the OrderBySinkNode
* the entry point is a SourceNode and not a TableSourceNode
was:
Now that we can print an ExecPlan (ARROW-15016), we can investigate unexpected
outputs. With the following chunk of code:
{code:r}
mtcars %>%
arrow_table() %>%
select(mpg, wt, cyl) %>%
filter(mpg > 20) %>%
arrange(desc(wt)) %>%
head(3) %>%
show_exec_plan()
#> ExecPlan with 3 nodes:
#> 2:SinkNode{}
#> 1:ProjectNode{projection=[mpg, wt, cyl]}
#> 0:SourceNode{}
{code}
* FilterNode disappears when head/tail are involved +
* we do not have additional information regarding the OrderBySinkNode
* the entry point is a SourceNode and not a TableSourceNode
> [R] Investigate Nodes missing from ExecPlan print out for nested queries
> ------------------------------------------------------------------------
>
> Key: ARROW-17184
> URL: https://issues.apache.org/jira/browse/ARROW-17184
> Project: Apache Arrow
> Issue Type: Bug
> Components: R
> Reporter: Dragoș Moldovan-Grünfeld
> Priority: Major
>
> *Update:* We currently do not print a plan for nested queries as that would
> involve the execution of the inner most query / queries. That is likely the
> reason why, in the chunk below, we do not see the filter and arrange nodes.
> ==========================
> *Original description:*
> Now that we can print an ExecPlan (ARROW-15016), we can investigate
> unexpected outputs. With the following chunk of code:
> {code:r}
> mtcars %>%
> arrow_table() %>%
> select(mpg, wt, cyl) %>%
> filter(mpg > 20) %>%
> arrange(desc(wt)) %>%
> head(3) %>%
> show_exec_plan()
> #> ExecPlan with 3 nodes:
> #> 2:SinkNode{}
> #> 1:ProjectNode{projection=[mpg, wt, cyl]}
> #> 0:SourceNode{}
> {code}
> * FilterNode disappears when head/tail are involved +
> * we do not have additional information regarding the OrderBySinkNode
> * the entry point is a SourceNode and not a TableSourceNode
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
