[
https://issues.apache.org/jira/browse/PHOENIX-1704?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14366545#comment-14366545
]
Serhiy Bilousov commented on PHOENIX-1704:
------------------------------------------
I have a question (mostly for my education) regarding
{code:title=YearFunction.java|borderStyle=solid}
// Some comments here
public boolean evaluate(Tuple tuple, ImmutableBytesWritable ptr) {
Expression expression = getChildExpression();
if (!expression.evaluate(tuple, ptr)) {
return false;
}
if ( ptr.getLength() == 0) {
return true; //means null
}
long dateTime = expression.getDataType().getCodec().decodeLong(ptr,
expression.getSortOrder());
DateTime dt = new DateTime(dateTime);
int year = dt.getYear();
PDataType returnType = getDataType();
byte[] byteValue = new byte[returnType.getByteSize()];
returnType.getCodec().encodeInt(year, byteValue, 0);
ptr.set(byteValue);
return true;
}
{code}
Long is 8 bytes but [UNSIGNED_TIMESTAMP
Type|http://phoenix.apache.org/language/datatypes.html#unsigned_timestamp_type]
is 12 bytes.
Will it possibly can cause some issues with [UNSIGNED_TIMESTAMP
Type|http://phoenix.apache.org/language/datatypes.html#unsigned_timestamp_type]?
Do we actually need to test real [UNSIGNED_TIMESTAMP
Type|http://phoenix.apache.org/language/datatypes.html#unsigned_timestamp_type]
(with nanos) ?
> Add year() built-in function
> ----------------------------
>
> Key: PHOENIX-1704
> URL: https://issues.apache.org/jira/browse/PHOENIX-1704
> Project: Phoenix
> Issue Type: Bug
> Reporter: Alicia Ying Shu
> Assignee: Alicia Ying Shu
> Fix For: 4.4
>
> Attachments: Phoenix-1704-v1.patch, Phoenix-1704-v2.patch,
> Phoenix-1704-v3.patch, Phoenix-1704-v4.patch, Phoenix-1704-v5.patch
>
>
> Support Year() with date and timestamp.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
