Help understanding code (with no comments) that uses for and switch

[cjohnman]

Can someone make some sense out of this code for me. I understand how the switch works but they lost me on the for loop using *fld[NUMFLD]. NUMFLD was defined as 10 and rec is a char array (rec[1000]) containing a line copied from a file. FLDSEP was defined as "|". I posted question on the pointer and it makes some sense now thanks to everyone who helped answer my questions. Now I am confused again. I can probably get thru id I knew what was happening during the for loop. Thanks for any assistance.

    920     int     fldcnt = 0;
    921     char    *fld[NUMFLD]; - *fld[10]
    922     int     error_flag = 0;
    923     char    tmprec[MAXRECLEN]; - tmprec[1000]
    924     int     index = 0;
    925
    926     sprintf(jnl.mesg,"%s",rec); 
    927     log_it(log_file,jnl);
    928
    929     strcpy(tmprec,rec); - copy the line into tmprec[1000]
    930     bzero(&fld,NUMFLD); 
    931     for (fld[fldcnt] = strtok(tmprec,FLDSEP); 
    932         fld[fldcnt] != NULL;
    933         fld[++fldcnt] = strtok(NULL,FLDSEP))
    934     {
    935         strip_trailing_spaces(fld[fldcnt]);
    936
    937         switch(fldcnt)
    938         {
    939             case 0:     /* bol */
    940                 pgi_ack_file.bol = atoi(fld[fldcnt]);
    941                 break;
    942             case 1:     /* delivery */
    943                 strcpy(pgi_ack_file.delivery,fld[fldcnt]);
    944                 break;
    945             case 2:     /* delivery_item */
    946                 pgi_ack_file.delivery_item = atoi(fld[fldcnt]);
    947                 break;
    948             case 3:     /* mfg_std */
    949                 pgi_ack_file.mfg_std = atoi(fld[fldcnt]);
    950                 break;
    951             case 4:     /* num_skids */
    952                 pgi_ack_file.num_skids = atoi(fld[fldcnt]);
    953                 break;
    954             case 5:     /* skid_num */
    955                 pgi_ack_file.skid_num = atoi(fld[fldcnt]);
    956                 break;
    957             case 6:     /* unit */
    958                 strcpy(pgi_ack_file.unit_txt,fld[fldcnt]);
    959                 break;
    960             case 7:     /* gross weight */
    961                 pgi_ack_file.skid_gross_wght_lbs = atof(fld[fldcnt]);
    962                 break;
    963             case 8:     /* net weight */
    964                 pgi_ack_file.skid_net_wght_lbs = atof(fld[fldcnt]);
    965                 break;
    966             case 9:     /* storage location */
    967                 pgi_ack_file.storage_location = atoi(fld[fldcnt]);
    968                 break;
    969             default:
    970                 strcpy(jnl.mesg,"Too many fields in record");
    971                 log_it(log_file,jnl);
    972                 sprintf(jnl.mesg,"Bad fldcnt %d. should be %d",fldcnt,NUMFLD);
    973                 log_it(log_file,jnl);
    974                 error_flag = 1;
    975                 break;
    976         }
    977
    978         if (error_flag == 1) return (IMCS_DB_FATAL);
    979     }
    980
    981     if (fldcnt != NUMFLD)
    982     {
    983         strcpy(jnl.mesg,"Too few fields in record");
    984         log_it(log_file,jnl);
    985         sprintf(jnl.mesg,"Bad fldcnt %d. should be %d",fldcnt,NUMFLD);
    986         log_it(log_file,jnl);
    987         error_flag = 1;
    988         return (IMCS_DB_FATAL);
    989     }

[matsp]

Interesting use of for-loop, essentially, it loops through all the fields in the tmprec, incrementing fldcnt when getting the "next" one.

     for (fld[fldcnt] = strtok(tmprec,FLDSEP); 
         fld[fldcnt] != NULL;
         fld[++fldcnt] = strtok(NULL,FLDSEP))

I personally would probably have written it like this:

     for (fldcnt = 0, fld[fldcnt] = strtok(tmprec,FLDSEP); 
         fld[fldcnt] != NULL;
         fld[fldcnt++] = strtok(NULL,FLDSEP))

to make sure it's clear fldcnt is starting at zero when you begin.


The switch is simply using the fldcnt to take the separated strings into their respective fields in a struct, using atoi(), atof() and strcpy() to convert/copy the value itself.

--
Mats

[cjohnman]

for (fld[fldcnt] = strtok(tmprec,FLDSEP);

I believe this is setting *fld[0] = first token in the first line that was stored in temprec[1000] and from here (somehow) it will continue storing the remaining 10 tokens into *fld[0...10]. I looked at one of the files and sure enough there are 10 "fields" all seperated by |. Hope this helps someone help me with the rest. Thanks.

-Carl

[cjohnman]

Thanks for taking time to explain that to me. I posted right after you did so I guess I am on the right path.

-Carl

[cjohnman]

fld[fldcnt++] = strtok(NULL,FLDSEP))

What is happening here? Thanks.

-Carl

[matsp]

Actually, it should only fill in 0..9 of fld[], and fldcnt should be 10 when the loop exits.

The code is actually broken in the case where the incoming data has more than 9 '|' in the line, it will overflow the fld[] array, so a more correct function would be:

for (fldcnt = 0, fld[fldcnt] = strtok(tmprec,FLDSEP); 
         fldcnt < 10 && fld[fldcnt] != NULL;
         fld[++fldcnt] = strtok(NULL,FLDSEP))

or

for (fldcnt = 0, fld[fldcnt] = strtok(tmprec,FLDSEP); 
         !errorflag && fld[fldcnt] != NULL;
         fld[fldcnt++] = strtok(NULL,FLDSEP))

--
Mats

[matsp]

fld[fldcnt++] = strtok(NULL,FLDSEP))

What is happening here? Thanks.

-Carl

strtok looks something like this internally:

char *strtok(char *str, int sep)
{
    static char *oldstr;
    char *orig = str; 
    if (str == NULL) str = oldstr;
    while(str && *str != sep) str++;
    if (!str) {
       oldstr = NULL;
       return NULL;
    }
    oldstr = str +1;  // Point to 
    return orig;
}

So if you pass in NULL, it will use the previous string after the current separotor.

--
Mats

[cjohnman]

Thank you. I will now move on to the struct. Fun stuff. Who knows maybe in 6 months with the help of this site and my books I may get to be a average programmer. NOT.

-Carl