>>>>>If the table is a dbf and not too big:

>>>>>lcData=Filetostr('timep.dbf')
>>>>>lnCount=Occurs('|',lcData)

>>>>>
>>>>>>I have 250 records in a database called timep.
>>>>>>Each record has 100 fields called p1,p2,p3 ..... p100
>>>>>>
>>>>>>Is there a global search i can do for a tally of the numbers of times the symbol "|" appears in the whole database - without having to count through each field of each record - ie without using scan or for-endfor. ie one or two lines of code.
>>>>>>
>>>>>>thanks for any suggestions
>>>>>>k
>>>>
>>>>
>>>>Maybe I'm wrong, but is this will accurate if the value "|" (0x7C or 124 decimal) appear in .DBF header?
>>>>
>>>>I think this code below will prevent that case
>>>>
>>>>

>>>>lcData=FILETOSTR("timep.dbf")
>>>>
>>>>*!* skip until first record
>>>>lnFirstRecordOffset = CTOBIN(SUBSTR(m.lcData, 9, 2), "2RS")
>>>>
>>>>lnCount=OCCURS("|", SUBSTR(m.lcData, m.lnFirstRecordOffset + 1))
>>>>

>>>
>>>Yes, I know, but karen wanted a maximum of two lines of code. :-)
>>
>>
>>As Sergey reply, your sample and my addition can't solve the real problem, in case if any numeric field / value in the records.
>>:D
>
>My first idea was something like

>Select * from timep.dbf into file somereallystupidname.txt plain
>lcData=filetostr('somereallystupidname.txt')
>lnCount=occurs('|',lcData)
>eras somereallystupidname.txt

>But that would leave the table open, and consume four lines of code (five if you close the table afterwards). However, in this case the client wanted maximum two lines of codes, and it's always the client who defines the conditions. :-)

If defined conditions did not solve their (client) real problem,
then as developer/consultant we should suggest the best method, right?
:)

BTW, your last sample looks like good one.
A text file will only contains real characters, not binary (except CR/LF).
This is good.