  | 
»  | 
 | 
  
 | 
 | 
There are four  SQL views added for TurboIMAGE/XL per DBEnvironment for the G.1 version.
These are the   SQL views   for TurboIMAGE/XL hash indices on key items and search items and for TPIs: SYSTEM.IMAGEKEY and CATALOG.IMAGEKEY |    |  
 The DBA can view all TurboIMAGE/XL hash indices associated with a database by examining
SYSTEM.IMAGEKEY. Other users can view the TurboIMAGE/XL hash indices to which they have  access by examining CATALOG.IMAGEKEY. For example, the DBA can issue this command: 
   isql => select * from system.imagekey where owner='music';
 
    -----------------------------------------------------
    INDEXNAME        | TABLENAME   | OWNER | UNIQUE | ...
    -----------------------------------------------------
    ALBUMCODE_M1     | ALBUMS      | MUSIC |  1     | ...
    COMPOSERNAME_M1  | COMPOSERS   | MUSIC |  1     | ...
    SELECTIONNAME_A1 | SELECTIONS_A| MUSIC |  1     | ...
    ALBUMCODE_D1     | SELECTIONS  | MUSIC |  0     | ...
    SELECTIONNAME_D2 | SELECTIONS  | MUSIC |  0     | ...
    COMPOSERNAME_D3  | SELECTIONS  | MUSIC |  0     | ...
    ALBUMCODE_D1     | LOG         | MUSIC |  0     | ...
    SELECTIONNAME_D2 | LOG         | MUSIC |  0     | ...
 |  
 Columns in SYSTEM.IMAGEKEY and CATALOG.IMAGEKEYThe following columns exist in both SYSTEM.IMAGEKEY and
CATALOG.IMAGEKEY  (PRIMARIES, SCCCOUNT, and DCCCOUNT are not
calculated or used in G.2): - INDEXNAME 
 Name of the TurboIMAGE/XL item, plus a suffix.
 The following suffixes are used by IMAGESQL:
 _M1 is used when entering definition for hash index on manual master
key item. _A1 is used when entering definition for hash index on automatic master
key item. _Dn is used when entering definitions for hash indices on
detail data set search items, where n is the path number from  1 to 16.
 - TABLENAME
 Name of the TurboIMAGE/XL data set on which the key is defined.
 - OWNER
 
Name of the TurboIMAGE/XL database (or the owner name used during
the ATTACH) on which the key is defined.
 - UNIQUE
 Uniqueness indicator: 0 if duplicates are allowed, that is, the index
is not unique. 1 if duplicates are not allowed, that is, the
key is unique. Hash index on the key item of a master data set (both automatic and
manual) is always unique, except when defined on P and Z
(decimal) data types. Hash indices on search
items of detail data sets are always non-unique.
 - NUMC
 Number of columns in the index.  NUMC is always 1. - COLNUMS
 A vector of 16 SYSTEM.COLUMN entries, which identifies the
column numbers that make up the key. In ISQL, each column number
is displayed as a field of 4 hexadecimal digits.
 - NDISTINCT
 Number of distinct key values. - PRIMARIES
 Number of primary slots used.
 - SCCCOUNT
 Synonym Chain Cluster Count, which is a measure of how well the
data is clustered on pages as the synonym chain (also known as the
secondary chain) is traversed.
 - DCCCOUNT
 
Detail Chain Cluster Count, which is a measure of how well the
data is clustered on pages as the detail chain is traversed. 
 SYSTEM.TPINDEX and CATALOG.TPINDEX |    |  
 The DBA can view all TPIs associated with a database by
examining  SYSTEM.TPINDEX. Other users can view the TPIs to which they have access by  examining CATALOG.TPINDEX. For example, the
DBA can issue: 
   isql => select * from system.tpindex where owner='music';
    ------------------------------------------------
    INDEXNAME    | TABLENAME  | OWNER | UNIQUE | ...
    ------------------------------------------------
    ACODE_T1     | ALBUMS     | MUSIC |  0     | ...
    ALBUMC_T2    | SELECTIONS | MUSIC |  0     | ...
    SELECTNAME_T3| SELECTIONS | MUSIC |  0     | ...
    COMPOSER_T4  | COMPOSERS  | MUSIC |  0     | ...
 |  
 Columns in SYSTEM.TPINDEX and CATALOG.TPINDEXThe following columns exist in both SYSTEM.TPINDEX and
CATALOG.TPINDEX  (NPAGES, NLEVELS, NLEAVES, NDISTINCT, NFIRST,
NPERKEY,  and CCOUNT   are not calculated or used in G.2):
 - INDEXNAME 
 Name of the TPI.  The following suffix is used by IMAGESQL when
registering TPIs: _Tn (Where n can be 1 to 400 depending on
the TPIs that exist on the database.)
 - TABLENAME 
 Name of the data set on which the TPI is defined.
 - OWNER
 Name of the TurboIMAGE/XL database (or the owner name used during
the ATTACH) on which the TPI is defined.
 - UNIQUE
 Uniqueness indicator: 0 if duplicates are allowed, that is, the index is not unique. 1 if duplicates are not allowed, that is, the index is unique.
 - CLUSTER
 Clustering indicator: 0 if the index is not a clustering index. 1 if the index is a clustering index. TPIs are always registered as non-clustering.
 - NUMC
 Number of columns in the index. - COLNUMS
 A vector of 16 SYSTEM.COLUMN entries, which identifies the
column numbers that make up the index. In ISQL, each column number is displayed as a field of 4
hexadecimal digits.
 - NPAGES
 Number of pages containing the index.
 - NLEVELS
 Number of levels in the TPI. - NLEAVES
 Number of leaf pages in the TPI.
 - NDISTINCT
 Number of distinct key values.
 - NFIRST
 Number of distinct first key values.
 - NPERKEY
 Number of pages per key. - CCOUNT
 Cluster count, which indicates how well the data of the index are
sorted: 0     before first UPDATE STATISTICS statement is processed. n (efficiency of clustering) best clustering if n=NPAGES of table
indexed; worst if n=NROWS of table indexed.
 - CTIME
 Time of creation: yyyymmddhhsstt.
 - COLDIRS
 A vector of 16 direction entries, which indicates the
direction of the corresponding column in the index definition.
In ISQL, each column number is displayed as a field of 4
hexadecimal digits. 5 Ascending. 6 Descending. 
 SYSTEM.INDEX and CATALOG.INDEX |    |  
 When you attach a database for which you have created one or more B-Tree indices on the key items of the master data sets, the ATTACH command enters definitions for the B-Tree indices on  the key items and their related search items of the detail data sets.  See the chapter, "B-Tree Indices," in the TurboIMAGE/XL Database Management System Reference Manual for more information on how to create B-Tree indices. The definitions of B-Tree indices on key items, except P and Z types, of the master data sets are entered 
as UNIQUE  indices.  For the related search items of ALL of the related 
detail data sets, the definitions  are entered as NON-UNIQUE  indices. View  your new B-Tree indices in the existing views, SYSTEM.INDEX and CATALOG.INDEX. For example: 
isql=> SELECT * FROM SYSTEM.INDEX WHERE OWNER = 'BTREE3';
Output will be truncated. (DBWARN 1)
select * from system.index where owner = 'BTREE3';                         
--------------------+--------------------+--------------------+------+-----
INDEXNAME           |TABLENAME           |OWNER               |UNIQUE|CLUST
--------------------+--------------------+--------------------+------+-----
TESTNAMEP16_B1      |PACKED              |BTREE3              |     0|
TESTNAMELENR2_B1    |REAL                |BTREE3              |     1|
TESTNAMELENR2_V1    |RDET1               |BTREE3              |     0|
TESTNAMELENR2_V1    |RDET2               |BTREE3              |     0|
TESTNAMELENR2_V1    |RDET3               |BTREE3              |     0|
TESTNAMEP16_V1      |PRZDET              |BTREE3              |     0|
TESTNAMELENR2_V2    |PRZDET              |BTREE3              |     0|
  |  
 From the above example, you can see that there are two master data sets,
PACKED and REAL, with B-Tree indices created on their key items.  The remaining are related detail data sets. Columns in SYSTEM.INDEX and CATALOG.INDEXThe following columns are in both SYSTEM.INDEX and CATALOG.INDEX: - INDEXNAME
 Name of the TurboIMAGE/XL item, plus a suffix. The following suffixes are used by IMAGE/SQL:
 - _B1  
 is used when entering the definition for B-Tree index on the master  key item.
 - _Vn  
 is used when entering the definition for B-Tree index on the related  search item of detail data set, where n is the path number from 1 to 16. 
  
 
All other column names are the same as for SYSTEM.TPINDEX and CATALOG.TPINDEX.  Refer to "Columns in SYSTEM.TPINDEX and CATALOG.TPINDEX," earlier in this appendix.  
 |