[jira] [Commented] (PHOENIX-1287) Use the joni byte[] regex engine in place of j.u.regex

[Shuxiong Ye (JIRA)]

    [ 
https://issues.apache.org/jira/browse/PHOENIX-1287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14372461#comment-14372461
 ] 

Shuxiong Ye commented on PHOENIX-1287:
--------------------------------------

Hi [~jamestaylor],

I have a question about PVarchar.toObject. When we are trying to put a String 
"" into ImmutableBytesWritable, and turn it back to String like the following 
code, we get null back. Why not "" back here?

{code:java}
        Expression srcExp = LiteralExpression.newConstant("", TYPE);
        ImmutableBytesWritable ptr = new ImmutableBytesWritable();
        boolean eval = srcExp.evaluate(null, ptr);
        String result = (String)srcExp.getDataType().toObject(ptr);
        System.out.println(result); // result = null, here
        assertTrue("".equals(result));
{code}

In PVarchar.java,
{code:java}
  @Override
  public Object toObject(byte[] bytes, int offset, int length, PDataType 
actualType,
      SortOrder sortOrder, Integer maxLength, Integer scale) {
    if (!actualType.isCoercibleTo(this)) {
      throwConstraintViolationException(actualType, this);
    }
    if (length == 0) {  // <---- here to set it null when length is zero
      return null;
    }
    if (sortOrder == SortOrder.DESC) {
      bytes = SortOrder.invert(bytes, offset, length);
      offset = 0;
    }
    return Bytes.toString(bytes, offset, length); // <-- here we call 
Bytes.toString from HBase, in which it will return "" when length is zero, and 
null when bytes ptr is null.
  }
{code}


> Use the joni byte[] regex engine in place of j.u.regex
> ------------------------------------------------------
>
>                 Key: PHOENIX-1287
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-1287
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: James Taylor
>            Assignee: Shuxiong Ye
>              Labels: gsoc2015
>
> See HBASE-11907. We'd get a 2x perf benefit plus it's driven off of byte[] 
> instead of strings.Thanks for the pointer, [~apurtell].



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)