[
https://issues.apache.org/jira/browse/PHOENIX-1142?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jeffrey Zhong updated PHOENIX-1142:
-----------------------------------
Attachment: PHOENIX-1142.patch
The attached patch addressed following issues:
1) Make TO_DATE function & internal data time parsing logic same. Both places
can recognize:
a) Phoenix previous default date format: "yyyy-MM-dd HH:mm:ss"
b) Make "HH:MM:ss" part optional as reported in JIRA PHOENIX-1274
c) All ISO 8601 compliant formats
For example BY DEFAULT following date time strings are able to be recognized:
"2015-01-27T16:17:57+00:00"
"2015-01-27T16:17:57Z"
"2015-01-27T16:17:59"
"2015-027"
"2015-W05"
"2015-W05-2"
"2015-01-27"
"2015-01-27 16:17:58"
"2015-01-27 16:17:58.123"
2) A user can directly pass date time string into a UPSERT statement without
using TO_DATE function to cast as following(like other sql)
UPSERT INTO <TABLE> values (..., "2013-01-01", "2015-01-27 16:17:58",...)
3) Fix the current JIRA issue so that CsvBulkLoad tool can use configuration
"phoenix.query.dateFormat" or "phoenix.query.dateFormatTimeZone" if CSV file
date time string isn't ISO8601 complaint.(Check
CsvBulkLoadToolIT#testBasicImport for example)
> Improve CsvBulkLoadTool to parse different Date formats
> -------------------------------------------------------
>
> Key: PHOENIX-1142
> URL: https://issues.apache.org/jira/browse/PHOENIX-1142
> Project: Phoenix
> Issue Type: Improvement
> Reporter: David Kjerrumgaard
> Assignee: Jeffrey Zhong
> Attachments: PHOENIX-1142.patch
>
>
> Currently, the CsvBulkLoadTool uses a single 'default' format to parse dates
> from the source CSV file. This can sometimes cause issues when the date
> fields don't match this default format.
> Would it be possible to add a method for specifying / overriding the
> 'default' date format?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
