[
https://issues.apache.org/jira/browse/LUCENE-4539?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13491020#comment-13491020
]
Robert Muir commented on LUCENE-4539:
-------------------------------------
I think you are probably right... I just got the straightforward ones done.
I wanted to suggest a possibility that maybe we add PackedInts.Header to
encapsulate what it stores?
Today we have:
{code}
Reader getReader(DataInput in)
Reader getReaderNoHeader(DataInput in, Format format, int version, int
valueCount, int bitsPerValue)
{code}
And if you want to read in the header yourself, you are doing:
{code}
final int version = CodecUtil.checkHeader(in, CODEC_NAME, VERSION_START,
VERSION_CURRENT);
final int bitsPerValue = in.readVInt();
assert bitsPerValue > 0 && bitsPerValue <= 64: "bitsPerValue=" + bitsPerValue;
final int valueCount = in.readVInt();
final Format format = Format.byId(in.readVInt());
{code}
So my idea would just be something like:
{code}
Reader getReader(DataInput in, Header header) {
return getReader(in, header.format, header.version, header.valueCount,
header.bitsPerValue);
}
{code}
and maybe
{code}
Header readHeader(DataInput in)
{code}
to encapsulate this stuff. we could still keep the "raw" versions around if we
wanted too.
I didnt want to overengineer anything, not sure if this would be useful outside
of this particular issue.
> DocValues impls should read all headers up-front instead of per-directsource
> ----------------------------------------------------------------------------
>
> Key: LUCENE-4539
> URL: https://issues.apache.org/jira/browse/LUCENE-4539
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/index
> Reporter: Robert Muir
> Attachments: LUCENE-4539.patch
>
>
> Currently, when DocValues opens, it just opens files. it doesnt read codec
> headers etc.
> Instead we read these every single time a directsource opens.
> I think it should work like PostingsReaders: e.g. the PackedInts impl would
> read its versioning info and codec headers and creating a new Direct impl
> should be a IndexInput.clone() + getDirectReaderNoHeader().
> Today its much more costly.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
