Better file handling in CL with the CLOSE command

CL is an often undervalued part of the i developer’s toolkit, and one that IBM have been quietly enhancing over the years. It does have some quirks however, the most annoying being that if you are reading a file, once you reach the end of the file there is no way of repositioning the pointer and reading it again.

Actually, there is.

The command you need to know about is the Close Database File (CLOSE) command:

The Close Database File (CLOSE) command will explicitly close a database file that was implicitly opened by a Receive File (RCVF) command run within the same CL program or ILE CL procedure. A file with the same open identifier must have been declared by a Declare File (DCLF) command within the same CL program or ILE CL procedure.

After the CLOSE command is run, the next RCVF command for this file will implicitly reopen the file and read a record from the file.

That last line is the really important piece. After you close the file, the next RCVF command will re-open it and your file pointer is back at the first record.

Here’s an example:

    dcl &FROMFILE *char 50
    dcl &TOMBR *char 50 value('/qsys.lib/LSCLIB.lib/IMPORTWORK.file/IMPORTWORK.mbr')
    dcl &QSH *char 50
    dclf file(FTPFILES) opnid(LIST)

    /* Merge Original FTP data into Reload File                               */
    ovrdbf file(INPUT) tofile(LSCLIB/QFTPSRC) mbr(GETFULL)
    ovrdbf FILE(OUTPUT) tofile(LSCLIB/QFTPSRC) mbr(FTPLOG)
    ftp ''
    dltovr OUTPUT
    dltovr INPUT
    callsubr MERGEDATA

    /* Merge correction FTP data into Reloqd File                             */
    ovrdbf file(INPUT) tofile(LSCLIB/QFTPSRC) mbr(GETCORR)
    ovrdbf FILE(OUTPUT) tofile(LSCLIB/QFTPSRC) mbr(FTPLOG)
    ftp ''
    dltovr OUTPUT
    dltovr INPUT
    callsubr MERGEDATA

/* Merge retrieved FTP records with missing data into Reload File             */

    /* List the retrieved files                                               */
    qsh cmd('ls /home/PAUL/FTPIN/ > /home/PAUL/SentFiles.txt')
    clrpfm FTPFILES
    cpyfrmstmf fromstmf('/home/PAUL/SentFiles.txt') +

    /* Then use this list to read through and import the files                */
    dowhile '1'
        rcvf opnid(LIST)
        monmsg msgid(CPF0864) exec(leave)

        /* Copy and process the stream file                                   */
        chgvar &FROMFILE value('/home/PAUL/FTPIN/' *tcat &LIST_FTPFILES)
        cpyfrmstmf fromstmf(&FROMFILE) tombr(&TOMBR) mbropt(*REPLACE)
        cpyf fromfile(LSCLIB/IMPORTWORK) tofile(LSCLIB/IMPORT) mbropt(*REPLACE) +
        call LSCMRG001R parm(&LIST_FTPFILES)

        /* Delete the stream file once we've finished with it                 */
        chgvar &QSH value('rm /home/PAUL/FTPIN/' *tcat &LIST_FTPFILES)
        qsh cmd(&QSH)


    /* Clean up and return                                                    */
    close opnid(LIST)
    qsh cmd('rm /home/PAUL/SentFiles.txt')



What I am doing here is retrieving a set of files from a customer’s FTP server, and then retrieving the corrections from the same server. After retrieving the files, I execute the MERGEDATA subroutine. This builds a list of the retrieved files and copies it to the FTPFILES physical file.

The DOWHILE loop copies each stream file into a physical file (IMPORT) and then calls program LSCMRG001R to merge the retrieved data into the enterprise application.

Note that once I leave the loop, I close the file. This means that I can reuse the MERGEDATA subroutine as many times as I need to, for whatever selection of files I happen to retrieve.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.