Daylight Summer School 2000, June 7-9, Santa Fe, NM

Daylight Worksheet - Configuring a Merlin pool -- WITH HINTS

Configuring a Merlin pool means specifying which datafields of which datatypes are loaded by the Merlinserver. This is done by editing the datatypes database. Specifically, the _P specifier for the datatype.

This exercise will show you one way to modify the pool configuration, and observe the effect on pool size.

  1. Use a database containing the CR datatype. The usual _P (Pool Inclusion) specifier is:
    Start the xvthor application and use the File -> Show all DBs button to display and then open-writeable the datatypes database of choice.

    % xvthor &

    Try not to edit the same database as someone else!

  2. Before modifying the pool, first load the pool and record the number of "bytes for pool" reported by the log file. You may try searching this field to confirm that it is unavailable. Then release the pool.

  3. Select Identifier Datatype to be ROOT DATATYPE and look up CR.

    Notice that the RIGHT mouse button must be used to open a menu with these XView applications. After typing in "CR", hit the return key to lookup the datatype TDT.

  4. Edit the _P specifier so it is now *;*;* (add one asterisk). This instructs the Merlinserver to load the third datafield as well as the first two. Save the edited datatree.

    Once the TDT widget is in modify mode, select the Pool Inclusion datafield by pressing the left mouse button. This should bring up the tdt-editing widget.

    Add the asterisk required and press Apply. Then press Save on the TDT widget to write the change to the database.

  5. Again load the pool and examine the log file to see the change in size. Search this field using xvmerlin to verify that it is now available.
    % more $DY_MERLIN_LOG_FILE
    look for medchem98demo...
    Mon Apr 12 13:36:48 1999: auto-opening
    loading datatypes: 23 datatype definitions loaded
    loading primary data:
    pool loaded: 1999 TDTs read, 1999 rows in pool
    Creating hash table...
    Scanning datatypes...
             0 dataitems of type SIMILARITY
          1999 dataitems of type $ROWID
          2397 dataitems of type PCN
             0 dataitems of type $GRF
          1607 dataitems of type CL
           970 dataitems of type PKA
          3485 dataitems of type $NAM
          4965 dataitems of type P
            22 dataitems of type PKA1
             0 dataitems of type $ISM
             0 dataitems of type ISM
          1999 dataitems of type $SMI
            18 dataitems of type MR
          2081 dataitems of type $WLN
             0 dataitems of type AMW
             9 dataitems of type P2
          1772 dataitems of type AC
            10 dataitems of type REM
           582 dataitems of type P1
          1825 dataitems of type $D3D
          1945 dataitems of type CR
          1993 dataitems of type $CAS
          1998 dataitems of type FP
          1998 dataitems of type F
          1827 dataitems of type CP
         33502 dataitems total
    Figuring out which indirect references to load ...
    Loading indirect data:
    1894 indirect-reference TDTs scanned, 1726 loaded
            (95667 bytes of raw indirect data, 151634 bytes total)
    Pool Information:
        662223 bytes for raw data
        263250 bytes for pool
         10012 bytes for hash table
         95667 bytes for indirect data
       1031152 bytes for pool.
        185168 bytes system overhead
       1216320 bytes total memory

Daylight Chemical Information Systems Inc.