[
https://issues.apache.org/jira/browse/SPARK-61023?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hyukjin Kwon updated SPARK-61023:
---------------------------------
Description:
In PySpark, the {{subset}} of {{df.dropna(how, thresh, subset)}} and
{{df.na.drop}} is a column name or a list or tuple of them, and {{how}} is
"any" or "all", else {{VALUE_NOT_ALLOWED}}. {{df.fillna(value, subset)}} takes
a bool, a dict, a float, an int or a string as the value, else
{{NOT_EXPECTED_TYPE}}. {{df.describe(*cols)}} takes column names or a list of
them. {{df.corr(col1, col2, method)}} takes only the method "pearson", else
{{VALUE_NOT_ALLOWED}}. {{df.approxQuantile(col, probabilities, relativeError)}}
takes a column or a list or tuple of them, and raises {{NEGATIVE_VALUE}} for a
negative relativeError.
The Rust-backed Python client's {{df.dropna(subset="x")}} and
{{df.na.drop(subset="x")}} raise {{TypeError: argument 'subset': Can't extract
`str` to `Vec`}}. {{df.dropna(how="bogus")}} is accepted, {{df.fillna([1])}}
fills the string 'List([Long(1)])', and {{df.describe(["x", "v"])}} raises a
TypeError. {{df.corr("x", "v", "kendall")}} returns the Pearson correlation,
and {{df.approxQuantile("v", [0.5], -1.0)}} fails on the server with an
IllegalArgumentException. {{df.approxQuantile(["x", "v"], [0.5], 0.0)}}, of a
list of columns, raises a TypeError. Also, the statistic functions {{cov}},
{{corr}}, {{crosstab}}, {{freqItems}}, {{approxQuantile}} and {{summary}} raise
TypeErrors rather than PySpark's {{NOT_EXPECTED_TYPE}} for arguments of other
types. Take the arguments as PySpark does.
was:
In PySpark, the {{subset}} of {{df.dropna(how, thresh, subset)}} and
{{df.na.drop}} is a column name or a list or tuple of them, and {{how}} is
"any" or "all", else {{VALUE_NOT_ALLOWED}}. {{df.fillna(value, subset)}} takes
a bool, a dict, a float, an int or a string as the value, else
{{NOT_EXPECTED_TYPE}}. {{df.describe(*cols)}} takes column names or a list of
them. {{df.corr(col1, col2, method)}} takes only the method "pearson", else
{{VALUE_NOT_ALLOWED}}. {{df.approxQuantile(col, probabilities, relativeError)}}
raises {{NEGATIVE_VALUE}} for a negative relativeError.
The Rust-backed Python client's {{df.dropna(subset="x")}} and
{{df.na.drop(subset="x")}} raise {{TypeError: argument 'subset': Can't extract
`str` to `Vec`}}. {{df.dropna(how="bogus")}} is accepted, {{df.fillna([1])}}
fills the string 'List([Long(1)])', and {{df.describe(["x", "v"])}} raises a
TypeError. {{df.corr("x", "v", "kendall")}} returns the Pearson correlation,
and {{df.approxQuantile("v", [0.5], -1.0)}} fails on the server with an
IllegalArgumentException. Take the arguments as PySpark does.
> Take the arguments of dropna, na.drop, fillna, describe, corr and
> approxQuantile in the Python client as PySpark does
> ---------------------------------------------------------------------------------------------------------------------
>
> Key: SPARK-61023
> URL: https://issues.apache.org/jira/browse/SPARK-61023
> Project: Spark
> Issue Type: Bug
> Components: Connect
> Affects Versions: connect-rust-4.3.0
> Reporter: Hyukjin Kwon
> Priority: Major
> Fix For: connect-rust-4.3.0
>
>
> In PySpark, the {{subset}} of {{df.dropna(how, thresh, subset)}} and
> {{df.na.drop}} is a column name or a list or tuple of them, and {{how}} is
> "any" or "all", else {{VALUE_NOT_ALLOWED}}. {{df.fillna(value, subset)}}
> takes a bool, a dict, a float, an int or a string as the value, else
> {{NOT_EXPECTED_TYPE}}. {{df.describe(*cols)}} takes column names or a list of
> them. {{df.corr(col1, col2, method)}} takes only the method "pearson", else
> {{VALUE_NOT_ALLOWED}}. {{df.approxQuantile(col, probabilities,
> relativeError)}} takes a column or a list or tuple of them, and raises
> {{NEGATIVE_VALUE}} for a negative relativeError.
> The Rust-backed Python client's {{df.dropna(subset="x")}} and
> {{df.na.drop(subset="x")}} raise {{TypeError: argument 'subset': Can't
> extract `str` to `Vec`}}. {{df.dropna(how="bogus")}} is accepted,
> {{df.fillna([1])}} fills the string 'List([Long(1)])', and
> {{df.describe(["x", "v"])}} raises a TypeError. {{df.corr("x", "v",
> "kendall")}} returns the Pearson correlation, and {{df.approxQuantile("v",
> [0.5], -1.0)}} fails on the server with an IllegalArgumentException.
> {{df.approxQuantile(["x", "v"], [0.5], 0.0)}}, of a list of columns, raises a
> TypeError. Also, the statistic functions {{cov}}, {{corr}}, {{crosstab}},
> {{freqItems}}, {{approxQuantile}} and {{summary}} raise TypeErrors rather
> than PySpark's {{NOT_EXPECTED_TYPE}} for arguments of other types. Take the
> arguments as PySpark does.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
