[
https://issues.apache.org/jira/browse/NIFI-2842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15548248#comment-15548248
]
Bojan Kostic commented on NIFI-2842:
------------------------------------
I just checked InferAvroSchema processor with large files. And i can conform
this issue.
Also i tried to change Number Of Records To Analyze field and it was still
failing.
I think that this option is broken, because i tried to set it to 2k even my
file have 1k lines.
> Would like InferAvroSchema and ConvertCSVToAvro to handle numbers better
> ------------------------------------------------------------------------
>
> Key: NIFI-2842
> URL: https://issues.apache.org/jira/browse/NIFI-2842
> Project: Apache NiFi
> Issue Type: New Feature
> Reporter: David Hicks
>
> Assume the following CSV:
> field1,field2
> 10,17
> 7,18.4
> InferAvroSchema will parse both as an integer field, but ConvertCSVtoAvro
> will explode when trying to convert the second line, because it's a float.
> One recommendation would be to allow the user to specify multiple lines and
> choose the least restrictive one. If it parses a field as an integer and
> then a double, the double will override the integer.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
