I'm not sure what happen, but all of a sudden we have spaces in our variables that we can't remove.

I've tried both TRIM_ and TRUNCATE and there is still a space there, but the default is '', no space at all, so where is this space coming from?

Code:

-SET &CLIENTPATH = '';
-TYPE CLIENTPATH = >>&CLIENTPATH<<
-SET &CLIENTPATH = TRIM_(BOTH, ' ', &CLIENTPATH);
-TYPE CLIENTPATH = >>&CLIENTPATH<<
-SET &CLIENTPATH = TRUNCATE(&CLIENTPATH);
-TYPE CLIENTPATH = >>&CLIENTPATH<<

Response:

CLIENTPATH = >> <<
CLIENTPATH = >> <<
CLIENTPATH = >> <<

This message has been edited. Last edited by: GavinL,

Is it a space ?

Can you try using UFMT to check.

Hi GavinL,
I don't think so, it is a space because when i apply ARGLEN function on the variable it is giving zero length.

-SET &CLIENTPATH ='';
-TYPE CLIENTPATH = >>&CLIENTPATH<<

-SET &LENGTH=ARGLEN(50, &CLIENTPATH, 'I2');
-TYPE &LENGTH

Thanks

WebFCOUS
8.1.04

My understanding is that variables always have at least one character, zero length values are not supported.

Well, I just did another test and I guess WebFOCUS evaluates spaces with blanks.. This day an age, not able to evaluate spaces. What odd results..

-SET &CLIENTPATH = '';
-SET &MC = &CLIENTPATH.LENGTH;
-SET &MV = IF &CLIENTPATH EQ ' ' THEN 'YES' ELSE 'NO';
-SET &MV2 = IF &CLIENTPATH EQ '' THEN 'YES' ELSE 'NO';
-TYPE MC = >>&MC<<
-TYPE MV = >>&MV<<
-TYPE MV2 = >>&MV2<<
-TYPE CLIENTPATH = >>&CLIENTPATH<<
-SET &CLIENTPATH = TRIM_(BOTH, ' ', &CLIENTPATH);
-TYPE CLIENTPATH = >>&CLIENTPATH<<
-SET &CLIENTPATH = TRUNCATE(&CLIENTPATH);
-TYPE CLIENTPATH = >>&CLIENTPATH<<
-SET &STROUT = STRREP (&CLIENTPATH.LENGTH, &CLIENTPATH, 1, ' ', 0, '', &CLIENTPATH.LENGTH, 'A&CLIENTPATH.LENGTH');
-TYPE STROUT = >>&STROUT<<

Results:

 MC = >>01<<
 MV = >>YES<<
 MV2 = >>YES<<
 CLIENTPATH = >> <<
 CLIENTPATH = >> <<
 CLIENTPATH = >> <<
 STROUT = >> <<

Here's something interesting:

-SET &MYVAR0 = ;
-SET &MYVAR1 = '';
-SET &MYVAR2 = ' ';
-SET &MYVAR3 = '  ';
-TYPE MYVAR0 LENGTH: &MYVAR0.LENGTH
-TYPE MYVAR1 LENGTH: &MYVAR1.LENGTH
-TYPE MYVAR2 LENGTH: &MYVAR2.LENGTH
-TYPE MYVAR3 LENGTH: &MYVAR3.LENGTH

&MYVAR0 has a length of zero. I don't know if this code is valid.

Interesting in deed! I think you found an undocumented feature.

Even set to _FOC_NULL returns 09. You found the only thing that returns what is expected.

Francesco, add this

-TYPE MYVAR0 LENGTH: &MYVAR0.LENGTH &MYVAR0.EXIST
-TYPE MYVAR1 LENGTH: &MYVAR1.LENGTH &MYVAR1.EXIST

and you'll see that myvar0 doesn't exist, but myvar1 does
so I use that feature to nuke some &vars that I've been messing with in a fex to make sure they don't overstay their welcome.
-SET &myvar0 = ;
and its gone

"I see", said the blind man as he fell over the cliff...

Yes, we use -SET &var =; to effectively null the variable.

In fact its a way to stop prompting if you don't want to use -DEFAULTH

quote:

Originally posted by Francis Mariani:
"I see", said the blind man as he fell over the cliff...

Miracle cure!

Have you considered CTRAN?