[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628369#comment-16628369
]
Marco Gaido commented on SPARK-25538:
-
Please do not use Blocker and Critical when reporting issues as they are
reserved for committer. Though, I agree this should be a blocker for 2.4.0 as
it is a correctness issue. cc [~cloud_fan]
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628388#comment-16628388
]
Wenchen Fan commented on SPARK-25538:
-
Hi [~Steven Rand], is it possible to upload the data file? We need to be able
to reproduce it in order to debug it, thanks! I'll set it as a blocker once I
confirm this is a correctness issue.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628449#comment-16628449
]
Steven Rand commented on SPARK-25538:
-
Hi [~cloud_fan], I'm still trying to create a scrubbed version of the file.
It's proving to be difficult since mutating the original DataFrame often
prevents the issue from reproducing (e.g., the example in the description where
calling {{withColumnRenamed}} before {{distinct.count}} leads to the correct
result being printed).
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628744#comment-16628744
]
Wenchen Fan commented on SPARK-25538:
-
cc [~kiszk] as well
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16629281#comment-16629281
]
Kazuaki Ishizaki commented on SPARK-25538:
--
Hi [~Steven Rand], would it be possible to share the schema of this DataFrame?
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16629561#comment-16629561
]
Steven Rand commented on SPARK-25538:
-
[~kiszk], yes, the schema is:
{code}
scala> spark.read.parquet("hdfs:///data").printSchema
root
|-- col_0: string (nullable = true)
|-- col_1: timestamp (nullable = true)
|-- col_2: string (nullable = true)
|-- col_3: timestamp (nullable = true)
|-- col_4: string (nullable = true)
|-- col_5: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_6: string (nullable = true)
|-- col_7: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_8: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_9: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_10: string (nullable = true)
|-- col_11: timestamp (nullable = true)
|-- col_12: integer (nullable = true)
|-- col_13: boolean (nullable = true)
|-- col_14: decimal(38,18) (nullable = true)
|-- col_15: long (nullable = true)
|-- col_16: string (nullable = true)
|-- col_17: integer (nullable = true)
|-- col_18: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_19: string (nullable = true)
|-- col_20: string (nullable = true)
|-- col_21: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_22: string (nullable = true)
|-- col_23: array (nullable = true)
||-- element: timestamp (containsNull = true)
|-- col_24: string (nullable = true)
|-- col_25: string (nullable = true)
|-- col_26: string (nullable = true)
|-- col_27: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_28: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_29: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_30: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_31: decimal(38,18) (nullable = true)
|-- col_32: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_33: string (nullable = true)
|-- col_34: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_35: decimal(38,18) (nullable = true)
|-- col_36: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_37: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_38: decimal(38,18) (nullable = true)
|-- col_39: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_40: string (nullable = true)
|-- col_41: string (nullable = true)
|-- col_42: string (nullable = true)
|-- col_43: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_44: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_45: string (nullable = true)
|-- col_46: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_47: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_48: string (nullable = true)
|-- col_49: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_50: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_51: array (nullable = true)
||-- element: string (containsNull = true)
|-- col_52: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
|-- col_53: string (nullable = true)
|-- col_54: decimal(38,18) (nullable = true)
|-- col_55: decimal(38,18) (nullable = true)
|-- col_56: decimal(38,18) (nullable = true)
|-- col_57: array (nullable = true)
||-- element: decimal(38,18) (containsNull = true)
{code}
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> wil
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16631605#comment-16631605
]
Kazuaki Ishizaki commented on SPARK-25538:
--
Thank for upload a schema. While I looked at the schema, I am still not sure
about the reason of this problem.
I would appreciate it if you could find a good input data that can reproduce a
problem.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16632729#comment-16632729
]
Steven Rand commented on SPARK-25538:
-
[~kiszk] that makes sense, I'll try to do so. The issue I've been having so far
is that when I run the UDF I've written to change the data (while preserving
number of duplicate rows), the resulting DataFrame doesn't reproduce the issue.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16633254#comment-16633254
]
Steven Rand commented on SPARK-25538:
-
[~kiszk] I've uploaded a tarball containing parquet files that reproduce the
issue but don't contain any of the values in the original dataset.
Specifically, some columns have been dropped, all strings have been changed to
"test_string", all values in col_50 have been changed to 0.0043, and the values
in col_14 have all been mapped from their original values to values between
0.001 and 0.0044.
This new DataFrame still reproduces issues similar to those in the description:
{code:java}
scala> df.distinct.count
res3: Long = 64
scala> df.sort("col_0").distinct.count
res4: Long = 73
scala> df.withColumnRenamed("col_0", "new").distinct.count
res5: Long = 63
{code}
I get those inconsistent/wrong results on {{2.4.0-rc2}} and if I check out
commit {{a7c19d9c21d59fd0109a7078c80b33d3da03fafd}}, which is SPARK-23713. If I
check out the commit immediately before, which is
{{fe2b7a4568d65a62da6e6eb00fff05f248b4332c}}, then all three commands return 63.
cc [~cloud_fan] – IMO this should block the 2.4.0 release.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16633568#comment-16633568
]
Kazuaki Ishizaki commented on SPARK-25538:
--
Thank you. I will check it tonight in Japan.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16634106#comment-16634106
]
Marco Gaido commented on SPARK-25538:
-
I was able to reproduce also using limit instead of sort:
{code}
scala> df.limit(80).distinct.count
res83: Long = 72
scala> df.distinct.count
res84: Long = 64
scala> df.limit(20).distinct.count
res88: Long = 20
scala> df.limit(20).distinct.collect.distinct.length
res89: Int = 17
{code}
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Major
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16634344#comment-16634344
]
Kazuaki Ishizaki commented on SPARK-25538:
--
This test case does not print {{63}}.
{code}
test("test2") {
val df = spark.read.parquet("file:///SPARK-25538-repro")
val c1 = df.distinct.count
val c2 = df.sort("col_0").distinct.count
val c3 = df.withColumnRenamed("col_0", "new").distinct.count
val c0 = df.count
print(s"c1=$c1, c2=$c2, c3=$c3, c0=$c0\n")
}
c1=64, c2=73, c3=64, c0=123
{code}
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Blocker
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16634430#comment-16634430
]
Dongjoon Hyun commented on SPARK-25538:
---
[~mgaido]'s PR, https://github.com/apache/spark/pull/22602, fixes the decimal
issue and looks reasonable.
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Blocker
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16634532#comment-16634532
]
Apache Spark commented on SPARK-25538:
--
User 'mgaido91' has created a pull request for this issue:
https://github.com/apache/spark/pull/22602
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Blocker
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16634534#comment-16634534
]
Apache Spark commented on SPARK-25538:
--
User 'mgaido91' has created a pull request for this issue:
https://github.com/apache/spark/pull/22602
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Priority: Blocker
> Labels: correctness
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-25538) incorrect row counts after distinct()
[
https://issues.apache.org/jira/browse/SPARK-25538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16637630#comment-16637630
]
Steven Rand commented on SPARK-25538:
-
Thanks all!
> incorrect row counts after distinct()
> -
>
> Key: SPARK-25538
> URL: https://issues.apache.org/jira/browse/SPARK-25538
> Project: Spark
> Issue Type: Bug
> Components: SQL
>Affects Versions: 2.4.0
> Environment: Reproduced on a Centos7 VM and from source in Intellij
> on OS X.
>Reporter: Steven Rand
>Assignee: Marco Gaido
>Priority: Blocker
> Labels: correctness
> Fix For: 2.4.0
>
> Attachments: SPARK-25538-repro.tgz
>
>
> It appears that {{df.distinct.count}} can return incorrect values after
> SPARK-23713. It's possible that other operations are affected as well;
> {{distinct}} just happens to be the one that we noticed. I believe that this
> issue was introduced by SPARK-23713 because I can't reproduce it until that
> commit, and I've been able to reproduce it after that commit as well as with
> {{tags/v2.4.0-rc1}}.
> Below are example spark-shell sessions to illustrate the problem.
> Unfortunately the data used in these examples can't be uploaded to this Jira
> ticket. I'll try to create test data which also reproduces the issue, and
> will upload that if I'm able to do so.
> Example from Spark 2.3.1, which behaves correctly:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 115
> {code}
> Example from Spark 2.4.0-rc1, which returns different output:
> {code}
> scala> val df = spark.read.parquet("hdfs:///data")
> df: org.apache.spark.sql.DataFrame = []
> scala> df.count
> res0: Long = 123
> scala> df.distinct.count
> res1: Long = 116
> scala> df.sort("col_0").distinct.count
> res2: Long = 123
> scala> df.withColumnRenamed("col_0", "newName").distinct.count
> res3: Long = 115
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
