CONTROL CARD(S) .........1.........2.........3.........4.........5.........6.........7.........8 REPORT AREA=PXX,DUMPS=TRACE FUNCTION=REPORT AREA=PXX DUMPS=TRACE
This page of the report shows the following:
*** CMD= SELFR DBID= 001 TABLE= POL USER= $$$ $ID.CBSTS001(001) 000102 DATE= 1062003 TIME= 145837 *********** REQUEST SPECIFICATION ************** DATAVIEW ELEMENT: PO OFFSET IN RECORD: 00000 LENGTH: 00005 DATAVIEW ELEMENT: LI OFFSET IN RECORD: 00005 LENGTH: 00003 DATAVIEW ELEMENT: PN OFFSET IN RECORD: 00008 LENGTH: 00003 DATAVIEW ELEMENT: QTY OFFSET IN RECORD: 00011 LENGTH: 00003 DATAVIEW ELEMENT: UCOST OFFSET IN RECORD: 00014 LENGTH: 00005 WHERE: EQ Z N PO 0000 0005 V 0005 12345 F1F2F3F4F5 WHERE: A EQ C N PN 0000 0003 V 0003 XYZ E7E8E9 KEY DEF: POLI ( 000B ) TYPE: NATIVE ( D0 ) ROWS: 00002974 FIELDS= 00000.005 ; 00005.003 ; CARD = 000000010 ; 000000200 STATS: 012C 0000 00CC 0000 00C8 0014 00C8 0014 00C8 0014 00C8 0014 00C8 0014 KEY DEF: PN ( 000F ) TYPE: KEYINC=Y ( 40 ) ROWS: 00002974 FIELDS= 00008.003 ; CARD = 000000100 ; STATS: 012C 0000 00CC 0000 00C8 0014 00C8 0014 00C8 0014 00C8 0014 00C8 0014 PRI. KEY PN FROM= XYZ (IN HEX)= E7E8E9 TO= XYZ (IN HEX)= E7E8E9 SEC. KEY POLI FROM= 12345 (IN HEX)= F1F2F3F4F5000000 TO= 12345 (IN HEX)= F1F2F3F4F5FFFFFF **************** OPTIMIZATION ******************* CBSOR REASONS: 1-5: NNNNN 6: D 7-9: YNY 10-17: YYNYNNYN 18-21: NNNN TYPE OPTIMIZATION......: POPULATION DEPENDENT INDEX BUFFER REFERENCES: 00000013 **************** SELECTION ********************** TYPE SELECTION.........: *INDIRECT* DUE TO: O KEY INTERSECTION O KEY NOT AVAILABLE **************** ORDERING *********************** TEMPORARY INDEX NOT REQUIRED. **************** RETRIEVAL ********************** DATA RECORDS ACCESSED DUE TO: SELECTION ***END OF SELFR CMD FOR SET 00000001 CBS SET NUMBER 00000001 BEING RELEASED TOTAL NBR INDEX ENTRIES SCANNED 00000013 TOTAL NBR DATA AREA ROWS READ 00000001 TOTAL NBR ROWS ACCEPTED 00000001 ***END-OF-SET***
This page of the report displays the following information:
Request is made by CA Ideal program CBSTS001, version 001, system $ID, on line 000102.
The section displays the following information:
The program requests a DATAVIEW with elements PO, LI, PN, QTY and UCOST for table POL. The element name is followed by its offset in the record (not including RCE) and its length. The first element is at offset 0.
Displays the contents of the Selection Predicate from the Request Qualification Area. In this example, rows where PO=12345 and PN=XYZ are requested. For an explanation of the remaining fields, see the CA Datacom/DB Database and System Administration Guide.
Displays information for the keys defined in table POL. This group of fields is repeated for each key.
Keys POLI and PN are followed by the key ID (000B) and (000F).
For the POLI key, the type is NATIVE. For the PN key, the type is KEYINC=Y. Codes representing the key attributes, which are for internal use, are printed in parentheses after the key type.
The number of rows indexed by this key. The number is initialized by the LOAD or RETIX function and is maintained by adds and deletes to the table.
Field PO starts at offset zero in the record for a length of 5. The second field of the POLI key follows the semicolon for an offset of 5 and a length of 3. Field PN references the first and only field of key PN which is at offset 8 in the record for a length of 3.
Displays the cardinalities for the columns in the index. For example, if the key contains the branch and dept columns and the cardinalities displayed are 10 and 50, this means there are 10 branches and 50 combinations of branches and departments.
Cardinality is important in understanding the selection efficiency of an index. To continue this example, if there are 1,000 rows in the table and the values are evenly distributed, the search condition "WHERE BRANCH = x" would search 1,000 divided by the cardinality of 10 which is 100 rows. If "AND DEPT = y" is added, then only 1,000 divided by 50, or 20 rows, would be searched
Cardinality is initialized by the LOAD or RETIX function. It is not updated by normal table maintenance requests. Therefore, optimum selection efficiency could require using RETIX to keep this statistic current.
Statistical population counting for internal use only.
Displays the name of the primary key used in the search and the range of key values in both character and hexadecimal. In this example: the optimizer selected key PN as primary traversal key with the range of key character values xyz to xyz and hexadecimal values E7E8E9 to E7E8E9.
Displays the name of the secondary key used in the search and the range of key values in both character and hexadecimal. In this example: it also selected key POLI as a secondary traversal key with the range of key character values 12345 to 12345 and hexadecimal values F1F2F3F4F5000000 to F1F2F3F4F5FFFFFF.
The section displays the following information:
See the CA Datacom/DB Database and System Administration Guide to interpret CBSOR values for 1-5, 6, 7-9, 10-17, and 18-21.
The choice of which key to use is POPULATION DEPENDENT.
Indicates the type of optimization.
This example requires 13 index buffer references to determine that the previously shown key PN is best.
The section displays the following information at TYPE SELECTION: Selection is INDIRECT because no single key can be used to directly locate qualifying rows (KEY NOT AVAILABLE), even though KEY INTERSECTION is used.
The section displays information about use of a temporary index.
TEMPORARY INDEX NOT REQUIRED. For an explanation of the reasons that may be listed if a temporary index is required, see the CA Datacom/DB Database and System Administration Guide.
The section displays the following information:
Indicates rows were retrieved due to SELECTION criteria.
SELFR processing has completed for the designated set.
An explicit (SELPR) or implicit request to release the set was processed.
Number of index entries read in order to satisfy the SELxx requests for set.
Number of data area rows read in order to satisfy the SELxx requests for set.
Number of rows returned to requestor.
No more diagnostics for this set.
If this is a frequent request, defining a key containing both PO and PN would eliminate statistical population counting and index scanning since TYPE SELECTION would then be DIRECT.
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|