Date:         Thu, 24 Feb 2005 00:29:14 -0500
Reply-To:     Richard Ristow <wrristow@mindspring.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Richard Ristow <wrristow@mindspring.com>
Subject:      Re: Data Processing
Comments: To: Josh Johnson <joshjohnson88990@yahoo.co.uk>
In-Reply-To:  <20050223135212.71800.qmail@web86903.mail.ukl.yahoo.com>
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 08:52 AM 2/23/2005, Josh Johnson wrote:

>V1 is a string variable representing telephone numbers. What I need to
>do is:
>a. remove any non-numeric characters.
>then
>b. add zero to any case which its first character is not zero.
>then
>c. remove any cases with a length different than 10 or 11 characters.
>and last
>d. keep only cases which begin with '03' or '04'.

See the following, which is tested; SPSS draft output.
- It uses your revised test data
- V1 is your input, DESIRED is the output from your example; EDITED is
the computed output
- Variables EDIT_A, EDIT_B, EDIT_C and EDIT_D are for debugging; they
have no effect on the computation.
.............
List

Notes
|-------------------------|--------------------|
|Output Created           |24 Feb 05  00:23:57 |
|-------------------------|--------------------|

REC V1              DESIRED

   1 35#5687/589     0355687589
   2 4245896354      04245896354
   3 12458/7#4369    .
   4 3145215863      03145215863
   5 04525687458     04525687458
   6 5245896354      .
   7 32/4#5875436    03245875436
   8 03145215863     03145215863
   9 57856987542536  .
  10 14452987642536  .

Number of cases read:  10    Number of cases listed:  10


STRING EDITED EDIT_A EDIT_B EDIT_C EDIT_D (A15).

COMPUTE EDITED=V1.

*  First, remove leading blanks from the input. (You don't  .
*  say, but I think this is implied, or makes sense.)       .
COMPUTE EDITED = LTRIM(EDITED).

*  a. remove any non-numeric characters.                    .
*  This is the pain in the neck, in SPSS. There's no        .
*  function to search for a character NOT on a list.        .

COMPUTE #END_PT = LENGTH(RTRIM(EDITED)).
COMPUTE #INDEX  = 1.
LOOP #I = 1 TO 15.
-  DO IF   INDEX ('0123456789',SUBSTR(EDITED,#INDEX,1)) = 0.
.     COMPUTE SUBSTR(EDITED,#INDEX) = SUBSTR(EDITED,#INDEX+1).
.     COMPUTE #END_PT = #END_PT - 1.
-  ELSE.
.     COMPUTE #INDEX  = #INDEX  + 1.
-  END IF.
END LOOP IF #INDEX GT #END_PT.

COMPUTE EDIT_A = EDITED.

*  b. add zero to any case which its first character is not .
*     zero.                                                 .

IF (SUBSTR(EDITED,1,1) NE '0') EDITED = CONCAT('0',EDITED).

COMPUTE EDIT_B = EDITED.

*  c. remove any cases with a length different than 10 or 11.
*     characters.                                           .

IF (NOT ANY(LENGTH(RTRIM(EDITED)),10,11))  EDITED = '.'.
COMPUTE EDIT_C = EDITED.

*  d. keep only cases which begin with '03' or '04'"        .

IF (NOT ANY(SUBSTR(EDITED,1,2),'03','04')) EDITED = '.'.
COMPUTE EDIT_D = EDITED.

LIST REC V1 DESIRED EDITED.
List
Notes
|-------------------------|--------------------|
|Output Created           |24 Feb 05  00:23:59 |
|-------------------------|--------------------|

REC V1              DESIRED         EDITED

   1 35#5687/589     0355687589      0355687589
   2 4245896354      04245896354     04245896354
   3 12458/7#4369    .               .
   4 3145215863      03145215863     03145215863
   5 04525687458     04525687458     04525687458
   6 5245896354      .               .
   7 32/4#5875436    03245875436     03245875436
   8 03145215863     03145215863     03145215863
   9 57856987542536  .               .
  10 14452987642536  .               .

Number of cases read:  10    Number of cases listed:  10