SAS Safety Scalable Performance Data Server 4.5 Manual De Usuario

•

The average number of chunks for any value in the index.

•

The size of the idx file for the index. The idx file maintains the value segment lists and
bitmaps.

•

The number of garbage bytes in the idx file. This is space in the file that was thrown
away and cannot be reclaimed. Garbage bytes can result from deleting values, updating
values, or appending values.

•

The percentage of garbage bytes in the idx file.

The average number of chunks for a segment list is a good indicator of the fragmentation
level of the value segment lists. As this value increases, it can affect the query performance
for the index. If the average number of chunks exceeds 10, you should consider reorganizing
the index to consolidate the segment lists.

The number of garbage bytes and the percentage of garbage bytes indicate the amount of
unused disk space being consumed by the index. To conserve and consolidate disk space,
consider reorganizing the index. Reorganizing the index frees up disk space when the
garbage content is high.

Use the -dist option of ixutil -stats to get the index distribution for the index:

> ixutil -stats x -dsn test -libpath /spds -dist
SAS Scalable Performance Data Server 
   4.5(TS M0) Build(Feb 26 2009, 11:50:08)
Index File Utility
Copyright (c) 1996-2008 by SAS Institute Inc, Cary NC 27513 USA

Statistics for Index X:
---------------------------------------------------
+--segment_size         = 8192
+--n_segments_in_tbl    = 5
+--n_values_in_index    = 2
+--n_vdeleted_values    = 1
+--percent_vdeleted     = 33.33
+--n_seglist_values     = 2
+--n_seglist_chunks     = 3
+--avg_chunks_per_list  = 1.00
+--idx_file_bytes       = 13304
+--idx_garbage_bytes    = 4272
+--percent_idx_garbage  = 32.11

+--Distribution Stats for Non Unique Values
+----minimum segments for all values = 4
+----maximum segments for all values = 5
+----average segments of any value = 4
+----average percentage of segments of any value = 90.00

Ixutil completed successfully

The distribution statistics include the following:

•

The number of unique values in the index.

•

The number of non-unique values in the index.