[
https://issues.apache.org/jira/browse/CASSANDRA-172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710441#action_12710441
]
Jun Rao commented on CASSANDRA-172:
-----------------------------------
3. Within an SSTable, for a given (key,CF) pair, all columns are stored
contiguously in a segment. The column index is created roughly as follows.
Everytime the size of the accumulated columns reaches a threshold, a column
index entry is created, which has the column name and an offset within the
segment. I call this a (column) block index. In the code, instead of reading
all columns from an SSTable at once, I read the columns a block at a time,
using the column index. This is more efficient if the number of columns in an
SSTable is large (that's the key difference between this api and get_slice).
Depending on how many columns are needed in this api, multiple column blocks
will need to be fetched from a sinlge SSTable. Does that explain things?
> A improved and more general version of get_slice
> ------------------------------------------------
>
> Key: CASSANDRA-172
> URL: https://issues.apache.org/jira/browse/CASSANDRA-172
> Project: Cassandra
> Issue Type: New Feature
> Reporter: Jun Rao
> Assignee: Jun Rao
> Fix For: 0.4
>
> Attachments: get_slice_from.patchv1, get_slice_from.patchv2
>
>
> Today, get_slice has to scan through all columns in every memtable and
> sstable to get a slice of columns. This becomes inefficient when the number
> of columns in a row is large. We need a more efficient API.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
