You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@arrow.apache.org by GitBox <gi...@apache.org> on 2022/04/15 06:28:35 UTC
[GitHub] [arrow-datafusion] liukun4515 opened a new pull request, #2241: fix: find the right wider decimal datatype for comparison operation
liukun4515 opened a new pull request, #2241:
URL: https://github.com/apache/arrow-datafusion/pull/2241
# Which issue does this PR close?
<!--
We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes #123` indicates that this PR will close issue #123.
-->
Closes #2232
For decimal with the comparison operation, we use the error algorithm to calculate the precision/scale.
Before: max(p1,p2),max(s1,s2)
Now: max(38,max(s1, s2) + max(p1-s1, p2-s2)), max(s1, s2, 38)
# Rationale for this change
<!--
Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes.
-->
# What changes are included in this PR?
<!--
There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR.
-->
schema
```
CREATE EXTERNAL TABLE food (a DECIMAL(10,6), b double, c boolean) STORED AS CSV LOCATION 'aggregate_simple.csv';
```
Before this fix, we got the plan https://github.com/apache/arrow-datafusion/pull/2200#discussion_r850030108
After this fix, we got the right converted type in the filter expr
```
❯ explain select * from food where a >= CAST(0.00002 as Decimal(10,8));
+---------------+----------------------------------------------------------------------------------------------------------+
| plan_type | plan |
+---------------+----------------------------------------------------------------------------------------------------------+
| logical_plan | Projection: #food.a, #food.b, #food.c |
| | Filter: #food.a >= Decimal128(Some(2000),10,8) |
| | TableScan: food projection=Some([0, 1, 2]), partial_filters=[#food.a >= Decimal128(Some(2000),10,8)] |
| physical_plan | ProjectionExec: expr=[a@0 as a, b@1 as b, c@2 as c] |
| | CoalesceBatchesExec: target_batch_size=4096 |
| | FilterExec: CAST(a@0 AS Decimal(12, 8)) >= CAST(Some(2000),10,8 AS Decimal(12, 8)) |
| | RepartitionExec: partitioning=RoundRobinBatch(16) |
| | CsvExec: files=[aggregate_simple.csv], has_header=false, limit=None, projection=[a, b, c] |
| | |
+---------------+----------------------------------------------------------------------------------------------------------+
```
# Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be updated before approving the PR.
-->
<!--
If there are any breaking changes to public APIs, please add the `api change` label.
-->
cc @alamb
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb merged pull request #2241: fix: find the right wider decimal datatype for comparison operation
Posted by GitBox <gi...@apache.org>.
alamb merged PR #2241:
URL: https://github.com/apache/arrow-datafusion/pull/2241
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb commented on a diff in pull request #2241: fix: find the right wider decimal datatype for comparison operation
Posted by GitBox <gi...@apache.org>.
alamb commented on code in PR #2241:
URL: https://github.com/apache/arrow-datafusion/pull/2241#discussion_r851340950
##########
datafusion/physical-expr/src/coercion_rule/binary_rule.rs:
##########
@@ -178,15 +178,30 @@ fn get_comparison_common_decimal_type(
}
};
match (decimal_type, &other_decimal_type) {
- (DataType::Decimal(p1, s1), DataType::Decimal(p2, s2)) => {
- let new_precision = p1.max(p2);
- let new_scale = s1.max(s2);
- Some(DataType::Decimal(*new_precision, *new_scale))
+ (d1 @ DataType::Decimal(_, _), d2 @ DataType::Decimal(_, _)) => {
+ get_wider_decimal_type(d1, d2)
}
_ => None,
}
}
+// Find the winder decimal type when both types are decimal.
Review Comment:
```suggestion
// Returns a `DataType::Decimal` that can store any value from either
// `lhs_decimal_type` and `rhs_decimal_type`
```
##########
datafusion/physical-expr/src/coercion_rule/binary_rule.rs:
##########
@@ -178,15 +178,30 @@ fn get_comparison_common_decimal_type(
}
};
match (decimal_type, &other_decimal_type) {
- (DataType::Decimal(p1, s1), DataType::Decimal(p2, s2)) => {
- let new_precision = p1.max(p2);
- let new_scale = s1.max(s2);
- Some(DataType::Decimal(*new_precision, *new_scale))
+ (d1 @ DataType::Decimal(_, _), d2 @ DataType::Decimal(_, _)) => {
+ get_wider_decimal_type(d1, d2)
}
_ => None,
}
}
+// Find the winder decimal type when both types are decimal.
+// The result decimal type is (max(s1, s2) + max(p1-s1, p2-s2), max(s1, s2)).
+fn get_wider_decimal_type(
+ lhs_decimal_type: &DataType,
+ rhs_type: &DataType,
+) -> Option<DataType> {
+ match (lhs_decimal_type, rhs_type) {
+ (DataType::Decimal(p1, s1), DataType::Decimal(p2, s2)) => {
+ // max(s1, s2) + max(p1-s1, p2-s2), max(s1, s2)
+ let s = *s1.max(s2);
+ let range = (p1 - s1).max(p2 - s2);
Review Comment:
👍 I see that the precision includes the scale 👍
##########
datafusion/physical-expr/src/coercion_rule/binary_rule.rs:
##########
@@ -557,15 +571,17 @@ mod tests {
DataType::Float32,
DataType::Float64,
DataType::Decimal(38, 10),
+ DataType::Decimal(20, 8),
];
let result_types = [
DataType::Decimal(20, 3),
DataType::Decimal(20, 3),
DataType::Decimal(20, 3),
- DataType::Decimal(20, 3),
- DataType::Decimal(20, 7),
- DataType::Decimal(30, 15),
+ DataType::Decimal(23, 3),
+ DataType::Decimal(24, 7),
+ DataType::Decimal(32, 15),
DataType::Decimal(38, 10),
+ DataType::Decimal(25, 8),
Review Comment:
👍
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] liukun4515 commented on a diff in pull request #2241: fix: find the right wider decimal datatype for comparison operation
Posted by GitBox <gi...@apache.org>.
liukun4515 commented on code in PR #2241:
URL: https://github.com/apache/arrow-datafusion/pull/2241#discussion_r851569242
##########
datafusion/physical-expr/src/coercion_rule/binary_rule.rs:
##########
@@ -557,15 +571,17 @@ mod tests {
DataType::Float32,
DataType::Float64,
DataType::Decimal(38, 10),
+ DataType::Decimal(20, 8),
];
let result_types = [
DataType::Decimal(20, 3),
DataType::Decimal(20, 3),
DataType::Decimal(20, 3),
- DataType::Decimal(20, 3),
- DataType::Decimal(20, 7),
- DataType::Decimal(30, 15),
+ DataType::Decimal(23, 3),
+ DataType::Decimal(24, 7),
+ DataType::Decimal(32, 15),
DataType::Decimal(38, 10),
+ DataType::Decimal(25, 8),
Review Comment:
This is inspired by your comments https://github.com/apache/arrow-datafusion/pull/2200#discussion_r850030108
Thanks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org