>Being the detail oriented guy that I am (uh huh), I decided to automatically generate a data dictionary in my framework. The datadict is really two table, one containing table information and one containing field information. The problem I've run into is when trying to generate the table that contains field info. Basically, I attempted to do the following:
> scan
> get table name from datadict
> display structure to a file
> append that file to a temporary table
> get rid of the strucure header info
> add any extra info to the temp table, eg table name
> append the temp table into the main table
> endscan
>
>This works great with one exception. The file output from the 'display struct to file' isn't always the same! So, for example, since I'm assuming that the 'field type' will always start at column 30 this trashes everything.
>
>Is there an easier way to get structure info into a table?


Ive done the same thing, also using two tables, I generate the field table using AFIELDS() like this:


use db_list.751 && the table with the file listings
recnum = reccount()
dimension dblist(recnum)
anum = 1
scan for !empty(filename)
dblist(anum) = filename
anum = anum + 1
endscan
use
sele datadict
goto top
for q = 1 to recnum
select 0
fname = dblist(q)
use &fname
dec = afields(zfield)
use
select datadict
for x = 1 to dec
append blank
replace file with fname
replace field_name with zfield(x,1)
replace field_type with zfield(x,2)
replace field_len with zfield(x,3)
replace field_dec with zfield(x,4)
next
release all like z*
next
sort to datadict.751 on file,field_name


I'm sure there are plenty of ways to control the looping, perhaps better than this, but it does create the necessary data dictionary tables.

Hope this helps.

Steve