Eztrieve- Get substring. But Position is dynamic

rvshetttigar · New User Joined: 13 Oct 2009 Posts: 6 Location: US

Hi

I give a COBOL example..I need the same functionality in EZTRIEVE.

MYSUBSTRING = MYSTRING(POS, 4096) ..
That means I need 4096 bytes from position POS (a variable)
out of field MYSTRING.

POS variable can have any value during runtime.

dick scherrer · Posted: Tue Oct 13, 2009 7:25 pm

Hello,

Define the input/output areas as arrays of 1-byte entries and code a loop to move 4096 bytes from position n in the "input" to position 1 of the "output" and increment both arrays each time thru the loop.

Ajay Baghel · Posted: Tue Oct 13, 2009 8:23 pm

Hi Rvshettigar,

The below piece of code works for me:
Here is sample EZPLUS logic that searches a string for the location of a word.

rvshetttigar · New User Joined: 13 Oct 2009 Posts: 6 Location: US

Thanks Ajay and d.sch

Javier Rodríguez · New User Joined: 18 Mar 2010 Posts: 10 Location: Spain

There is a nice solution in Eztrieve to the string-substring problem:

dick scherrer · Posted: Fri Aug 13, 2010 8:31 pm

Hello,

Thank you for the suggestion

For this to be more useful for people with a similar requirement, but little Easytrieve experience, a more detailed explanation would help. . .

PeterHolland · Posted: Sat Aug 14, 2010 11:11 am

Javier Rodríguez · New User Joined: 18 Mar 2010 Posts: 10 Location: Spain

Hello Peter:

Note that my solution is different than yours.

I am redefining MYSTRING with MYSUBSTR, so I don't need any MOVE to obtain de substring I need in MYSUBSTR, except update addecuatelly POS.

MYSUBSTR is a mask overlapping MYSTRING.

Next code is part of a running program. For each input record (REGENT) I obtain in WCADENA, calling EXTRAE-CADENA appropiatelly, some substrings I need, substrings delimited by CRLF characters (the input record is a SWIFT MT message):

PeterHolland · Posted: Tue Aug 17, 2010 12:55 pm

Hello Javier,

Javier Rodríguez · New User Joined: 18 Mar 2010 Posts: 10 Location: Spain

Hello Peter:

I think you are mixing your solution with mine. Let me remember you that I have not defined any array (OCCURS clause) in my solution, so it makes no sense the calculation you are indicating ("index will be POS * (length of element)").

I think the Easytrieve Programmer Guide manual explains my technic in the paragraph Data Strings.

I suggest that you test my example. Of course don't doubt to ask me any question you could have.

Pleasure to discuss this item with you.

PeterHolland · Posted: Tue Aug 17, 2010 1:51 pm

Javier,

first im mixing nothing.
Your working program is indexing a 1-byte field so updating the index
will point to the next byte.
Your first example defines a 4096 byte length field, updating the index
will point to the next field of 4096 bytes.
I dont have to test your example, cause 1 byte indexing is child play.
On the other hand indexing fields larger than 1 byte is quite another
story.
To illustrate my point an excerpt from the manual :

Any data field definition can contain the INDEX attribute. An index can be used
to reference data fields which occur multiple times. If you do not use an index,
you must either use subscripts or assign individual field names to multiple field
occurrences.
The data field starting location is adjusted by the contents of its indexes to
determine the desired field occurrence. The INDEX indexname value is set to:
(desired occurrence number - 1) * (length of element)

Javier Rodríguez · New User Joined: 18 Mar 2010 Posts: 10 Location: Spain

Dear Peter, yes, you are mixing both solutions. Please don't get ungry with this affirmation. The only thing I pretend is to let you see the difference between your solution and mine.

The calculation of the index makes sense when you have elements, and you have elements when you have the OCCURS clause. You need to calculate the index when you need to reference the right element.

My working program uses, as well, a two-byte mask to scan the input file, indexed by the same index: PUNTERO-ENT. But I have used this technic extensively, with any needed length. For example, 4096. It works fine, if you don't go through the boundaries of the redefined field.

I hope to have been more clear! Sorry for my difficulties with english.

PeterHolland · Posted: Tue Aug 17, 2010 2:50 pm

Javier,

for the last time. If you update PUNTERO-ENT by 1 the internal ponter will point to the next 2 bytes of SCAN-EX2 cause the length is 2. Has nothing to do with an OCCURS clause, so if the length is 4096 an index update by 1 will result in the next 4096 bytes.
Thats how Easytrev works.

Javier Rodríguez · New User Joined: 18 Mar 2010 Posts: 10 Location: Spain

Great, Peter.

I agree you.

Nice to discuss with you... a pleasure!

Till next time.

PeterHolland · Posted: Tue Aug 17, 2010 2:56 pm

Javier,

thanks.

See you around.