SAS Scalable Performance Data Server 4.5, Administrator'S Manual

Page: 231 / 271

+--n_seglist_chunks = 2
+--avg_chunks_per_list = 1.00
+--idx_file_bytes = 9008
+--idx_garbage_bytes = 0
+--percent_idx_garbage = 0.00
Create a Join Index
Assume there are SPD Server tables in a domain in directory
/tmp
. A user has created two
tables, Table1 and Table2 that can be joined on column ID . An SPD Server index exists
on the column ID for both tables. A join index is created on the tables to allow a Parallel
Range Join on column ID.
Use the
-crejidx
option of the SPD Server ixutil command to create the join index.
> ixutil -crejidx Table1,ID Table2,ID
-libpath /tmp
-joinparts 4
SAS Scalable Performance Data Server
4.5(TS M0) Build(Feb 26 2009, 11:50:08)
Index File Utility
Copyright (c) 1996-2009 by SAS Institute Inc, Cary NC 27513 USA
Ixutil completed successfully.
Generate Join Index Statistics
Now, get statistics on the join index that you created, using the
-statjidx
option of the
ixutil command. The statistics are printed for each join range of the index, as well as for
the overall index. The range statistics identify each range (sobs=starting observation,
eobs=ending observation), the number of unique join keys that exist in the range, and the
number of keys that will be joined in the range for each table.
> ixutil -statjidx Table1,ID Table2,ID
-libpath /tmp
SAS Scalable Performance Data Server
4.5(TS M0) Build(Feb 26 2009, 11:50:08)
Index File Utility
Copyright (c) 1996-2009 by SAS Institute Inc, Cary NC 27513 USA
Stat of Join Index Table1.jdxid.table2.jdxid.0.0.0.spds9: Nranges=4
-----------------------------------------------------------------------
+-Range 0
+----<Table1,ID>: sobs=1 eobs=25000 (Sorted)
+-------unique_keys=25000, max_occurance=1
+-------obs=25000, joinobs=25000, rangepct=100.00
+----<Table2,ID>: sobs=1 eobs=10000 (Sorted)
+-------unique_keys=10000, max_occurance=1
+-------obs=10000, joinobs=10000, rangepct=100.00
+-Range 1
Generate Join Index Statistics
225