I'm new to WebFOCUS, and I could use a utility to map out TABLE FILE requests, with DEFINEs, JOINs, etc, so I can see how the data flows through a procedure.

I also want to search through all the files in an app directory.

I am working on something to help me do this, but do not want to "re-invent the wheel" so to speak.

Is there already something out there, or should I continue working on mine ?

Is there a need for something like this ?

Thanks in advance for your response !

This message has been edited. Last edited by: Kerry,

If you are in Dev Studio, open the procedure in the Dev Studio tool and select the Logical view tab at the bottom

In fact you have two questions.

We are working on the second for a while since we have maybe the same requirements .
What we do now is running a proces that starts with a DIR command on the DOS level, we pipe the output to a file and that file is the source for our internal administration.
What we would like to have, is the possibility to read in the fex and use the documentation.

I agree with Frank regarding question 2 and I do lots and lots of that. I have a generic master called LINE80 to which I filedef the results of the directory listing command. I parse the heck of the results and create reports for any number of reasons.

As for question 1, that is more like a FOCUS internals class. Since the FOCUS language is interpretive, data doesn't actually 'flow' but there are operations that are performed in a certain order. For instance, with or without a join:

1. Read the record.
2. Apply screening conditions against 'real' fields, i.e. those in the master.
3. Perform any defines.
4. Apply screening conditions against defined fields.
5. Build the internal matrix.
6. Process any COMPUTEs on the internal matrix.
7. Calculate any requested totals.
8. Present the report.

I'm sure that I have forgotten a thing or two but those are the 8 commandments I live by when deciding how to architect a program.

Just create a generic MFD (like Ginny uses) and FILEDEF the program itself - in other words you would write a program about what programs contain, EXECute, -INCLUDE, DEFINE, etc.

Since FEXs are just text, you could probably write a Perl script that would be much more robust and could do everything you're looking for but that may be beyond some people - might be a cool weekend project though...

Thanks RickW and GinnyJakes
- Being text files they can be more processed easily.
Since I was working on importing from SQL in MS Access, I wrote a routine that searches every program in an application directory and produces a text file with the filename and the line # where the search string was found, plus the complete line that was found

I'm working on a version that searches for WF-specific keywords, such as TABLE FILE, DEFINE, JOIN, etc, to produce a 'Data Flow' report for each file.

Both the 'Map' and 'Search' functions will recursively search all -INCLUDE files as well.

Thanks again for any input you may have !

Here is a sample output from my (MS Access VBA) 'Search' program :
(Indentations indicate recursive search into an INCLUDEd file)

 8/19/2008 @ 7:27:42 PM       Search for 'DEFINE' in C:\ibi\apps\USDA\
--------------------------------------------------
Searching 'a9xqkmj6-Main'
 4    C:\ibi\apps\USDA\a9xqkmj6-Main.fex       : DEFINE FILE 7GCOUNTY_PROGRESS
-INCLUDE a9xqkmj6.fex        <--- Processing INCLUDEd file here 
      4     C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 7GCOUNTY_PROGRESS
      33    C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 7GSTATE_PROGRESS
      136   C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 1PROGR[/COLOR]
 36   C:\ibi\apps\USDA\a9xqkmj6-Main.fex       : DEFINE FILE 7GSTATE_PROGRESS
 139  C:\ibi\apps\USDA\a9xqkmj6-Main.fex       : DEFINE FILE 1PROGR
--------------------------------------------------------------------------------
Searching 'a9xqkmj6'
 4    C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 7GCOUNTY_PROGRESS
 33   C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 7GSTATE_PROGRESS
 136  C:\ibi\apps\USDA\a9xqkmj6.fex               : DEFINE FILE 1PROGR
--------------------------------------------------------------------------------
Searching 'aaacarinst' <-- Text not found in this file
Searching 'afxyubdo'
 4    C:\ibi\apps\USDA\afxyubdo.fex               : DEFINE FILE 7GOFFICE_GOAL
--------------------------------------------------------------------------------
Searching 'bbjsx5cr'
 5    C:\ibi\apps\USDA\bbjsx5cr.fex               : DEFINE FILE 7GACRES
 32   C:\ibi\apps\USDA\bbjsx5cr.fex               : DEFINE FILE 7GPLAN
 296  C:\ibi\apps\USDA\bbjsx5cr.fex               : DEFINE FILE 1CN
--------------------------------------------------------------------------------
Searching 'bk6tbw0z'
 3    C:\ibi\apps\USDA\bk6tbw0z.fex               : DEFINE FILE 7GCOUNTY_PROGRESS
 72   C:\ibi\apps\USDA\bk6tbw0z.fex               : DEFINE FILE 1QPROG
 344  C:\ibi\apps\USDA\bk6tbw0z.fex               : DEFINE FILE 1GOAL7
 560  C:\ibi\apps\USDA\bk6tbw0z.fex               : DEFINE FILE 2HOLD
--------------------------------------------------------------------------------
Searching 'bpi_perf'
 5    C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 7GSTATE_GOAL_C
 35   C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 7GCOUNTY_PROGRESS
 64   C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 7GSTATE_PROGRESS
 81   C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 7GSTATE_GOAL_S
 377  C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 1REG3
 465  C:\ibi\apps\USDA\bpi_perf.fex               : DEFINE FILE 2REGACRO
-------------------------------------------------------------------------------- 

If you enter an empty string to search, the program will find "nothing," and simply return all of the files in a single composite document with code lines numbered (including INCLUDEd code, separately numbered)

(NOTE : The funky filenames were not my design !)

"Stop me if you've heard this one before !"

This message has been edited. Last edited by: Charlz,

That is one of the nice things of having plain old text files - all sorts of ways to to do it.

I concur Rick,

In fact a recent bulk update to all of our reports has been made exceedingly simple with tools like notepad++, winmerge and replacem which we used to identify and carry out our updates - i think it's a real strength of the code

as you say gives you all manner of options and methods...

I have written something similar.

It doesn't map the flow, but it does scan fex's and report on -INCLUDE's, Masters and Set Commands. It also report of a range of possible issues, such as missing DM labels and possible test remenants.

We use it as a check mechanism before deployment to the Test environments.

This code will list a load a focus db with every code line of every .fex found in the folder mentioned in &DIR2LIST. It make use of the APP query command. You can do the same kind of thing for all you master files using the CHECK FILE <master> HOLD get a hold file containing the details of the master file

Once you have these Data bases it is a simple thing to create reports parse the text line with the GETTOK, POSIT and SUBSTRING functions.

<code>
-SET &TEMPPATH = TEMPPATH (100,'A100');
-SET &FILESPEC = &TEMPPATH || 'FEXIN.MAS';
-SET &DIR2LIST = 'D:\IBI\APPS\IBINCCEN';
FILEDEF MAST1 DISK &FILESPEC
-RUN
-WRITE MAST1 FILE=FEXIN ,SUFFIX=FIX,$
-WRITE MAST1 SEGNAME=SEG01 ,SEGTYPE=S0,$
-WRITE MAST1 FIELD=TEXT, ,A80,A80V,$
-CLOSE MAST1
-SET &FILESPEC = &TEMPPATH || 'CODE_DB.MAS';
FILEDEF MAST2 DISK &FILESPEC
-RUN
-WRITE MAST2 FILE=CODE_DB ,SUFFIX=FOC,$
-WRITE MAST2 SEGNAME=SEG01 ,SEGTYPE=S1,$
-WRITE MAST2 FIELD=FILENAME, ,A80,$
-WRITE MAST2 SEGNAME=SEG02 ,SEGTYPE=S1,PARENT=SEG01,$
-WRITE MAST2 FIELD=LINE, ,I8,$
-WRITE MAST2 FIELD=TEXT, ,A80,$
-CLOSE MAST2
CREATE FILE CODE_DB
-RUN
APP MAP DIR2LIST &DIR2LIST
APP QUERY DIR2LIST HOLD
-RUN
TABLE FILE FOCAPPQ

PRINT FILENAME COMPUTE PAD/A64 = ' ';
WHERE FILENAME CONTAINS '.fex'
OR FILENAME CONTAINS '.FEX'
ON TABLE HOLD AS FILELIST FORMAT ALPHA
END
-RUN
-GETLOOP
-READ FILELIST NOCLOSE &FILEIN.A40.
-IF &IORETURN NE 0 THEN GOTO ENDLOOP;
-SET &FILESPEC = &DIR2LIST || '\' || &FILEIN;
FILEDEF FEXIN DISK &FILESPEC
-RUN
TABLE FILE FEXIN
PRINT
COMPUTE FILENAME/A80 = &FILEIN.QUOTEDSTRING;
COMPUTE LINE/I8 = LAST LINE + 1;
TEXT
ON TABLE HOLD AS HOLDFEX FORMAT ALPHA
END
-RUN
MODIFY FILE CODE_DB
FIXFORM FROM HOLDFEX
MATCH FILENAME
ON MATCH CONTINUE
ON NOMATCH INCLUDE
MATCH LINE
ON MATCH UPDATE TEXT
ON NOMATCH INCLUDE
DATA ON HOLDFEX
END
-RUN
-GOTO GETLOOP
-ENDLOOP
TABLE FILE CODE_DB
SUM TEXT
BY FILENAME
BY LINE
END
-RUN

</code>

Jim

This is pretty slick coding. I tried it on our database, but it is cutting off the first 2 character in the text.
the FILENAME and
COMPUTE LINE/I8 = LAST LINE + 1 is working fine
but in the TEXT is cut the data off. can you give me an educated guess on why this is happening, I have tried everything I can think of with no luck

Thanks
Chuck

quote:

-WRITE MAST1 FIELD=TEXT, ,A80,A80V,$

Try it without the V.

While we are on the subject of documentation, here's one I wrote a long time ago to print out .mas files. We had a need to print out .mas files, indexes and Oracle table names. (I wrote this back when FOCUS field names were limited to 12 characters and file and segment names were limited to 8.) But I still find it handy since I all the indexes on the Oracle table. It produces an OFFLINE report but could be changed to PDF without much trouble.

 APP PATH printmas
-*
-* Remove the existing output file from /opt/ibi/apps/reports
-* You may see an error message if output file is not generated yet
-*
-UNIX rm &&RPTPATH..masrpt.out
-*
-*  Copy .mas and .acx file to execution folder
-*
-UNIX cp /opt/ibi/apps/masters/*.mas .
-UNIX cp /opt/ibi/apps/prodacx/*.acx .
-*
-* Create list of .mas files and redirect the output to a new file
-* named mas.list
-*
-UNIX ls -A *.mas > mas.list
 SET MORE=OFF, MSG=OFF
 FILEDEF MASLIST DISK mas.list ( LRECL 32 RECFM V
-RUN
-*
-*  MASLOOP(Execution Loop) starts here
-*
-SET &LOOPCNT=0;
-*
-MASLOOP
-READ MASLIST NOCLOSE &MASNAME.A32.
-SET &LOOPCNT=&LOOPCNT+1;
-*-IF &LOOPCNT GT 50 THEN GOTO DONE;
-IF &IORETURN EQ 1 THEN GOTO DONE;
-IF &MASNAME EQ 'master.mas' THEN GOTO MASLOOP;
-IF &MASNAME EQ 'acx.mas' THEN GOTO MASLOOP;
-*
-UNIX cp &MASNAME master.ftm
-RUN
-*
-SET &UPMASNM=UPCASE(32,&MASNAME,'A32');
-SET &POS=POSIT(&MASNAME,32,'.',1,'I2');
-SET &POS=&POS+1;
-SET &ORANAME='N/A';
-SET &ACXNAME=OVRLAY(&MASNAME,32,'acx',3,&POS,'A32');
-*
-*Check the existence of .acx file for a corresponding .mas file
-*(retcode 0 is successful). If retcode is anything BUT 0 the control
-*will be passed to SKIPACX
-*
 STATE &ACXNAME
-RUN
-SET &NOACX=&RETCODE;
-*
-IF &NOACX NE 0 THEN GOTO SKIPACX;
-*
-*If retcode is 0 the excution will resume as normal within MASLOOP
-*
 FILEDEF ACX DISK &ACXNAME ( LRECL 80 RECFM V
-*
 DEFINE FILE ACX
   WKFILE1/A32=GETTOK(ACXREC,89,-1,'.',32,'A32');
   WKFILE2/A32=CTRAN(32,WKFILE1,44,32,'A32');
   FILENM/A32=UPCASE(32,WKFILE2,'A32');
 END
-*
-* Get the file name now. It will be saved to SAVE file
-*
 TABLE FILE ACX
 PRINT FILENM
    IF RECORDLIMIT EQ 1
    ON TABLE SAVE
 END
-*
? FILEDEF
 DEFINE FILE ACX CLEAR
 END
-*
 FILEDEF ACX  CLEAR
-RUN
-*
-* Read the ORACLE file name from SAVE file and clear the file
-*
-READ SAVE &ORANAME.A32.
-RUN
FILEDEF SAVE CLEAR
-RUN
-UNIX rm save.ftm
-*
-*  Collect the indexes for a given table.  Format and add as comments
-*  to the corresponding Master File Description
-*
-SET &QT=HEXBYT(39,'A1');
-SET &TESTORA=&QT | &ORANAME || &QT;
-*
 SET SQLENGINE=SQLORA
 SQL SET SERVER &&KIDS_SERVER
 SQL PREPARE LSTINDEX FOR
 SELECT INDEX_NAME, SUBSTR(COLUMN_NAME,1,42) COLNM
   FROM ALL_IND_COLUMNS
  WHERE TABLE_NAME = &TESTORA
  ORDER BY INDEX_NAME, COLUMN_POSITION ;
 END
-*
 OFFLINE
 FILEDEF OFFLINE DISK master.ftm ( LRECL 89 RECFM V APPEND
-*
 DEFINE FILE LSTINDEX
   INDX_LINE/A89='  ' | INDEX_NAME | '  ' | COLNM;
 END
-*
 TABLEF FILE LSTINDEX
 HEADING
" "
"======================================================================="
" "
"  INDEXES "
" "
  PRINT INDX_LINE AS ''
     BY INDEX_NAME NOPRINT
     ON INDEX_NAME SUBFOOT
" "
     ON TABLE SET PAGE NOPAGE
 END
-*
 DEFINE FILE LSTINDEX CLEAR
 END
-*
 OFFLINE CLOSE
-RUN
-*
-* The SKIPACX starts here ( this part of code executes if retcode is
-* anything BUT 0
-*
-SKIPACX
-*
 OFFLINE
 FILEDEF MASTER  DISK master.ftm ( LRECL 89 RECFM V
 FILEDEF OFFLINE DISK masrpt.out ( LRECL 90 RECFM V APPEND
-*
 TABLEF FILE MASTER
 HEADING
"PAGE:<TABPAGENO><60>RUN DATE: &DATEMDYY </1>"
 PRINT MASREC AS ''
 FOOTING BOTTOM
"ORACLE: &ORANAME <56>FOCUS: &UPMASNM"
 ON TABLE SET LINES 72
 END
-*
 OFFLINE CLOSE
 FILEDEF OFFLINE CLEAR
 FILEDEF MASTER CLEAR
-RUN
-UNIX rm master.ftm
-*
-* SKIPACX ends here and returns execution back to MASLOOP
-*
-GOTO MASLOOP
-*
-* DONE takes control of execution here( end of file). It will close the MASLIST file
-* and move output file (masrpt.out) to /opt/ibi/apps/reports folder
-*
-DONE
-CLOSE MASLIST
-RUN
-UNIX mv masrpt.out &&RPTPATH

The report looks like this.

 PAGE:    1                                               RUN DATE: 01/08/2008 
  
  FILE=KDSAPLY2, SUFFIX=SQLORA,                                                           
  $                                                                                       
  $ Contact Participants (Read Only)                                                      
  $                                                                                       
  SEGNAME=TCONTACT, SEGTYPE=S0,$                                                          
    FIELD=TA_CONTACT  , CONTACT_ID        , I8    , I4                ,$                  
    FIELD=TA_CCH_ID   , CL_COL_HHMEM_ID   , I8    , I4                ,$                  
    FIELD=TA_ROLECAT  , ROLE_TYP_CDE      , I5    , I4                ,$                  
    FIELD=TA_CCH_NAME , CL_COL_HHMEM_NME  , A80   , A80   , MISSING=ON,$                  
    FIELD=TA_UPDDTTM  , LST_UPDT_DT       , HYYMDS, HYYMDS            ,$                  
    FIELD=TA_UPDDT    , LST_UPDT_DT       , YYMD  , DATE              ,$                  
    FIELD=TA_UPDBY    , LST_UPDT_ID       , A10   , A10               ,$                  
                                                                                          
    =======================================================================               
                                                                                          
      INDEXES                                                                             
                                                                                          
      XCONTACT_APPLIES                CONTACT_ID                                          
      XCONTACT_APPLIES                CL_COL_HHMEM_ID                                     
      XCONTACT_APPLIES                ROLE_TYP_CDE                                        
                                                                                          
      XCONTACT_APPLIES1               ROLE_TYP_CDE                                        
      XCONTACT_APPLIES1               CL_COL_HHMEM_ID                                     
                                                                                          
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  ORACLE: TCONTACT_APPLIES                               FOCUS: KDSAPLY2.MAS              

Thanks Jack

I thought i had tried that yesterday, but maybe not, because that did the trick by removing the V

Sorry about the extra "V"

I did a copy and paste pull the key parts out of several procedures. Then quickly ran it to check for sytax flaws.

Jack Thank you

You're quite welcome.

Just thought I'd throw my 2¢ into this old thread.

I use the DOS findstr command and mapped drives (think it doesn't do UNC paths too well).

ex: x:\findstr /si "term" *.fex

/s - Subdirectories
/i - Case insensitive

Findstr also plays well with regular expressions.

ex: x:\findstr /si "\<term" *.fex

will find TERMinate but not exTERMinate.

"multi terms" does not work as you may think, it acts as an OR, not an AND. I have gotten around this by doing a 2 steps, one to hold and the second to scan the held files.

/m - just the filenames/paths (default is to show the line of context, which is sometimes quite helpful).

ex: 
x:\findstr /sim "term1" *.fex > c:\term1_out.txt
x:\findstr /si "term2" /f:"c:\term1_out.txt" > c:\term2_out.txt

You can also pass a text file fill of arguments (i.e. a selected list of terms).

x:\findstr /si /f:"c:\terms.txt" *.fex 

Read more here: http://ss64.com/nt/findstr.html

- ABT