Case to Vars with incremental vars?

[DKUKEC]

I am trying to use case to vars to do the following... any suggested syntax would be greatly appreciated. This is what have so far; however, the syntax does not work since I have duplicate code values within the same case_id. Thank you Damir ******************************************************* SORT CASES BY CASE_ID COURT_ORDER_DATEX CODESC. CASESTOVARS /ID=CASE_ID /INDEX = CODE /GROUPBY = VARIABLE / DROP COURT_ORDER CODE CODE_DESC. EXECUTE. CASE_ID COURT_ORDER CODE CODE_DESC 2005MHXXXXX 05-Jan-2005 1 Petition Involumtary Assessment 2005MHXXXXX 24-Jan-2005 6 Order (General) 2005MHXXXXX 04-Feb-2005 4 Order for Involuntary Treatment 2005MHXXXXX 23-Feb-2005 6 Order (General) 2005MHXXXXX 03-Mar-2005 6 Order (General) WOULD LIKE CASE_ID CODE_DATE1 CODE_DATE6.1 CODE_DATE4 CODE_DATE6.2 CODE_DATE6.3 2005MHXXXXX 05-Jan-2005 24-Jan-2005 04-Feb-2005 23-Feb-2005 03-Mar-2005

[Bruce Weaver]

Damir, your post came across as one uninterrupted string in the Nabble archive.  Here's what it was meant to look like, I think.

----

I am trying to use case to vars to do the following... any suggested syntax would be greatly appreciated.

This is what have so far; however, the syntax does not work since I have duplicate code values within the same case_id.

Thank you
Damir


*******************************************************

SORT CASES BY CASE_ID COURT_ORDER_DATEX CODESC.
CASESTOVARS
   /ID=CASE_ID
   /INDEX = CODE
   /GROUPBY = VARIABLE / DROP COURT_ORDER CODE CODE_DESC.
EXECUTE.

CASE_ID              COURT_ORDER     CODE   CODE_DESC
2005MHXXXXX 05-Jan-2005 1          Petition Involumtary Assessment
2005MHXXXXX 24-Jan-2005 6          Order (General)
2005MHXXXXX 04-Feb-2005 4          Order for Involuntary Treatment
2005MHXXXXX 23-Feb-2005 6          Order (General)
2005MHXXXXX 03-Mar-2005 6          Order (General)

WOULD LIKE

CASE_ID           CODE_DATE1   CODE_DATE6.1  CODE_DATE4   CODE_DATE6.2   CODE_DATE6.3
2005MHXXXXX   05-Jan-2005    24-Jan-2005    04-Feb-2005      23-Feb-2005       03-Mar-2005

[David Marso]

I would seem obvious that using CODE as the index might be counter productive?
See the archives for how to build a counter using LAG, then use that counter as INDEX.

Bruce Weaver wrote

Damir, your post came across as one uninterrupted string in the Nabble archive.  Here's what it was meant to look like, I think.

----

I am trying to use case to vars to do the following... any suggested syntax would be greatly appreciated.

This is what have so far; however, the syntax does not work since I have duplicate code values within the same case_id.

Thank you
Damir


*******************************************************

SORT CASES BY CASE_ID COURT_ORDER_DATEX CODESC.
CASESTOVARS
   /ID=CASE_ID
   /INDEX = CODE
   /GROUPBY = VARIABLE / DROP COURT_ORDER CODE CODE_DESC.
EXECUTE.

CASE_ID              COURT_ORDER     CODE   CODE_DESC
2005MHXXXXX 05-Jan-2005 1          Petition Involumtary Assessment
2005MHXXXXX 24-Jan-2005 6          Order (General)
2005MHXXXXX 04-Feb-2005 4          Order for Involuntary Treatment
2005MHXXXXX 23-Feb-2005 6          Order (General)
2005MHXXXXX 03-Mar-2005 6          Order (General)

WOULD LIKE

CASE_ID           CODE_DATE1   CODE_DATE6.1  CODE_DATE4   CODE_DATE6.2   CODE_DATE6.3
2005MHXXXXX   05-Jan-2005    24-Jan-2005    04-Feb-2005      23-Feb-2005       03-Mar-2005

[Andy W]

You can remake your "code" variable to conform to the new mappings, example below.

*****************************************************************************************************************.
data list list (",") / CASE_ID (A20) CODE_DATE (A20) CODE (A3) CODE_DESC (A50).
begin data
2005MHXXXXX,05-Jan-2005,1,Petition Involumtary Assessment
2005MHXXXXX,24-Jan-2005,6,Order (General)
2005MHXXXXX,04-Feb-2005,4,Order for Involuntary Treatment
2005MHXXXXX,23-Feb-2005,6,Order (General)
2005MHXXXXX,03-Mar-2005,6,Order (General)
end data.
dataset name orig.

*Make a new code variable that signifies order.
string CODE2 (A3).
sort cases by CASE_ID CODE CODE_DATE.
DO IF $casenum = 1. or CASE_ID <> lag(CASE_ID).
    compute #id = 1.
ELSE IF CASE_ID <> lag(CASE_ID) or CODE <> lag(CODE).
    compute #id = 1.
ELSE IF CASE_ID = lag(CASE_ID) and CODE = lag(CODE).
    compute #id = lag(#id) + 1.
END IF.
compute CODE2 = CONCAT(RTRIM(CODE),".",STRING(#id,F1.0)).
exe.

*Then usual casestovars.
SORT CASES by CASE_ID CODE2.
CASESTOVARS
/ID = CASE_ID
/INDEX = CODE2
/SEPERATOR = "_"
/GROUPBY=VARIABLE.
*****************************************************************************************************************.

If you don't want codes that have multiple instances to not be "?.1" you will need to do another step (e.g. AGGREGATE counts) to recode them.

[DKUKEC]

Thank you Bruce for cleaning up my post.  Can this be avoided in the future if I use text rather HTML?  

Thank you Andy, your proposed solution worked after a slight modification allowing for two decimals for the CODE2 var.

Cheers,
Damir

[Bruce Weaver]

"Can this be avoided in the future if I use text rather HTML?"

That's an empirical question.  Try it and see what happens!  ;-)

DKUKEC wrote

Thank you Bruce for cleaning up my post.  Can this be avoided in the future if I use text rather HTML?  

Thank you Andy, your proposed solution worked after a slight modification allowing for two decimals for the CODE2 var.

Cheers,
Damir

[Andy W]

Quick run-down for HTML, short answer is yes not clicking "Message is in HTML Format" should save formatting as it is visible in the text editing box on Nabble.

1) There is a button to preview the message, always do this regardless of HTML or not!
2) Paragraphs in HTML are marked by a "<p>" tag, and in xml spec. all tags need a closing tag, so the end of the paragraph would have a "</p>" tag. If you don't have these, the text will run all-together as in your post.
3) Useful extra, I like HTML for long code snippets, this can be accomplished by a "<pre><code>" tag, then copy-paste the whole snippet, then end with a "</code></pre>". I'll leave it to the interested to see what the code and the pre tag do.

[Rich Ulrich]

I wonder if the List is somehow randomly fouling up messages that
it relays ... either, in tiny ways, or differently to different readers.

The post that Bruce re-wrote was one that looked perfectly fine to me.
I'm reading what is sent by SPSS-L, not from Nabble.

I'm reading with Firefox, through MSN-Live mail.  On the other hand,
there was a post from Bruce a couple of days ago that was compressed/
wrapped around in the way that he described.  I have seen a few such
posts before.  Lately, I discovered that I could see them properly formatted
if I hit Reply (even though I don't intend to).  That also works to reformat
some posts that HTML and which first show themselves to me as spread
across many extra lines. 

One thing that further inspires my opening statement is that, just now,
I read a post from Ryan that is messed up, mainly by having a whole lot
of the letter a -- at the start of lines, and within lines.  That looks (I think)
like a bad translation of the characters that make borders.  "Reply" did
nothing to this one.

--
Rich Ulrich

> Date: Wed, 5 Jun 2013 12:40:55 -0700

> From: [hidden email]
> Subject: Re: Case to Vars with incremental vars?
> To: [hidden email]
>
> Quick run-down for HTML, short answer is yes not clicking "Message is in HTML
> Format" should save formatting as it is visible in the text editing box on
> Nabble.
>
> 1) There is a button to preview the message, always do this regardless of
> HTML or not!
> 2) Paragraphs in HTML are marked by a "<p>" tag, and in xml spec. all tags
> need a closing tag, so the end of the paragraph would have a "</p>" tag. If
> you don't have these, the text will run all-together as in your post.
> 3) Useful extra, I like HTML for long code snippets, this can be
> accomplished by a "<pre><code>" tag, then copy-paste the whole snippet, then
> end with a "</code></pre>". I'll leave it to the interested to see what the
> code and the pre tag do.
>
>
>
> -----
> Andy W

...

[Bruce Weaver]

Re Rich's point about Reply, Damir's original compressed post looked fine when I clicked Reply in Nabble.  

FWIW, Nabble has a check box for "Message is in HTML Format", but I never check it.

Rich Ulrich-2 wrote

I wonder if the List is somehow randomly fouling up messages that
it relays ... either, in tiny ways, or differently to different readers.

The post that Bruce re-wrote was one that looked perfectly fine to me.
I'm reading what is sent by SPSS-L, not from Nabble.

I'm reading with Firefox, through MSN-Live mail.  On the other hand,
there was a post from Bruce a couple of days ago that was compressed/
wrapped around in the way that he described.  I have seen a few such
posts before.  Lately, I discovered that I could see them properly formatted
if I hit Reply (even though I don't intend to).  That also works to reformat
some posts that HTML and which first show themselves to me as spread
across many extra lines.  

One thing that further inspires my opening statement is that, just now,
I read a post from Ryan that is messed up, mainly by having a whole lot
of the letter a -- at the start of lines, and within lines.  That looks (I think)
like a bad translation of the characters that make borders.  "Reply" did
nothing to this one.

--
Rich Ulrich



> Date: Wed, 5 Jun 2013 12:40:55 -0700
> From: [hidden email]
> Subject: Re: Case to Vars with incremental vars?
> To: [hidden email]
>
> Quick run-down for HTML, short answer is yes not clicking "Message is in HTML
> Format" should save formatting as it is visible in the text editing box on
> Nabble.
>
> 1) There is a button to preview the message, always do this regardless of
> HTML or not!
> 2) Paragraphs in HTML are marked by a "<p>" tag, and in xml spec. all tags
> need a closing tag, so the end of the paragraph would have a "</p>" tag. If
> you don't have these, the text will run all-together as in your post.
> 3) Useful extra, I like HTML for long code snippets, this can be
> accomplished by a "<pre><code>" tag, then copy-paste the whole snippet, then
> end with a "</code></pre>". I'll leave it to the interested to see what the
> code and the pre tag do.
>
>
>
> -----
> Andy W
...