[jira] [Commented] (FLINK-4794) partition_by_hash() crashes if no parameter is provided
[ https://issues.apache.org/jira/browse/FLINK-4794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15566067#comment-15566067 ] Yakov Goldberg commented on FLINK-4794: --- So need to add proper error message to partition_by_hash() > partition_by_hash() crashes if no parameter is provided > --- > > Key: FLINK-4794 > URL: https://issues.apache.org/jira/browse/FLINK-4794 > Project: Flink > Issue Type: Bug > Components: Python API >Reporter: Yakov Goldberg > > partition_by_hash() crashes if no parameter is provided. > Looks like a line of code was missed, check distinct() > def distinct(self, *fields): > f = None > if len(fields) == 0: > f = lambda x: (x,) > fields = (0,) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-4794) partition_by_hash() crashes if no parameter is provided
[ https://issues.apache.org/jira/browse/FLINK-4794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15566057#comment-15566057 ] Chesnay Schepler commented on FLINK-4794: - that is a bug and should be changed. > partition_by_hash() crashes if no parameter is provided > --- > > Key: FLINK-4794 > URL: https://issues.apache.org/jira/browse/FLINK-4794 > Project: Flink > Issue Type: Bug > Components: Python API >Reporter: Yakov Goldberg > > partition_by_hash() crashes if no parameter is provided. > Looks like a line of code was missed, check distinct() > def distinct(self, *fields): > f = None > if len(fields) == 0: > f = lambda x: (x,) > fields = (0,) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-4794) partition_by_hash() crashes if no parameter is provided
[ https://issues.apache.org/jira/browse/FLINK-4794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15566054#comment-15566054 ] Yakov Goldberg commented on FLINK-4794: --- Error message when using partition_by_hash() with no parameter Traceback (most recent call last): File "/tmp/flink_plan-299738636/plan.py", line 382, in dd24 \ File "/tmp/flink_plan-299738636/flink/plan/DataSet.py", line 505, in partition_by_hash if isinstance(fields[0], TYPES.FunctionType): IndexError: tuple index out of range > partition_by_hash() crashes if no parameter is provided > --- > > Key: FLINK-4794 > URL: https://issues.apache.org/jira/browse/FLINK-4794 > Project: Flink > Issue Type: Bug > Components: Python API >Reporter: Yakov Goldberg > > partition_by_hash() crashes if no parameter is provided. > Looks like a line of code was missed, check distinct() > def distinct(self, *fields): > f = None > if len(fields) == 0: > f = lambda x: (x,) > fields = (0,) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-4794) partition_by_hash() crashes if no parameter is provided
[ https://issues.apache.org/jira/browse/FLINK-4794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15566036#comment-15566036 ] Yakov Goldberg commented on FLINK-4794: --- Going to add error message. But in distinct(), if no parameter is passed, parameter enforced to be 0. > partition_by_hash() crashes if no parameter is provided > --- > > Key: FLINK-4794 > URL: https://issues.apache.org/jira/browse/FLINK-4794 > Project: Flink > Issue Type: Bug > Components: Python API >Reporter: Yakov Goldberg > > partition_by_hash() crashes if no parameter is provided. > Looks like a line of code was missed, check distinct() > def distinct(self, *fields): > f = None > if len(fields) == 0: > f = lambda x: (x,) > fields = (0,) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-4794) partition_by_hash() crashes if no parameter is provided
[ https://issues.apache.org/jira/browse/FLINK-4794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15564802#comment-15564802 ] Chesnay Schepler commented on FLINK-4794: - This is intended, assuming the appropriate java exception is thrown. The error message would be helpful. There is no partition_by_hash function without parameters in the java & scala counterparts, if you don't pass a parameter they will throw an exception. And this is good; just picking an (effectively arbitrary) field will in many cases lead to bad behavior. > partition_by_hash() crashes if no parameter is provided > --- > > Key: FLINK-4794 > URL: https://issues.apache.org/jira/browse/FLINK-4794 > Project: Flink > Issue Type: Bug > Components: Python API >Reporter: Yakov Goldberg > > partition_by_hash() crashes if no parameter is provided. > Looks like a line of code was missed, check distinct() > def distinct(self, *fields): > f = None > if len(fields) == 0: > f = lambda x: (x,) > fields = (0,) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
