Deleting Cases

[Nyougo Omae.]

Hello,

I would like to delete some cases from my data set using syntax.
Kindly help.

Nyougo.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Hal 9000]

select if not ( ----state conditions defining cases you want to delete-----).

On 10/17/07, Nyougo Omae. <[hidden email]> wrote:

> Hello,
>
> I would like to delete some cases from my data set using syntax.
> Kindly help.
>
> Nyougo.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Hal 9000]

In that case, you could use the $casenum function:

select if $casenum <> 1.
exe.

Just remember that after case #1 is gone, the first next case that
isn't deleted becomes case 1. It's better to use an id that is tied
directly to the case:

select if ~(id = '123456').

* (~ = "not").

- or if you want to delete by some combination of classifications:

select if ~(sex = 'M' & hair_color = 'Brown').

*this deletes males with brown hair.

Whatever you do, backup your original dataset before you start playing
around with select statements. It's a real good way to lose a lot of
data fast!

I much prefer to create a flag using the conditions I plan to use to
visually verify that I got it right. So, for example:

compute flag = ~(sex = 'M' & hair_color = 'Brown').
exe.

Then visually confirm that the right people have been flagged for
deletion. After you're convinced, change the compute statment to a
select statement. But that's just me - I'm cautious.

Good Luck!
Gary


On 10/17/07, [hidden email] <[hidden email]> wrote:

>
> Hello,
>
> Just to make sure that i follow. Say i want to delete case 1
>
> select if not ( case nb = 1) . I am new to syntax :)
>
> Thank you.
>
>
> -------------- Original message --------------
> From: Hal 9000 <[hidden email]>
>
> > select if not ( ----state conditions defining cases you want to
> delete-----).
> >
> > On 10/17/07, Nyougo Omae. wrote:
> > > Hello,
> > >
> > > I would like to delete some cases from my data set using syntax.
> > > Kindly help.
> > >
> > > Nyougo.
> > >
> > > =====================
> > > To manage your subscription to SPSSX-L, send a message to
> > > [hidden email] (not to SPSSX-L), with no body text except the
> > > command. To leave the list, send the command
> > > SIGNOFF SPSSX-L
> > > For a list of commands to manage subscriptions, send the command
> > > INFO REFCARD
> > >
> >
> > =====================
> > To manage your subscription to SPSSX-L, send a message to
> > [hidden email] (not to SPSSX-L), with no body text except the
> > command. To leave the list, send the command
> > SIGNOFF SPSSX-L
> > For a list of commands to manage subscriptions, send the command
> > INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Richard Ristow]

At 06:22 PM 10/17/2007, Hal 9000 wrote:

>select if ~(id = '123456').
>
>* (~ = "not").
>
>- or if you want to delete by some combination of classifications:
>
>select if ~(sex = 'M' & hair_color = 'Brown').
>
>*this deletes males with brown hair.
>
>Whatever you do, backup your original dataset before you start playing
>around with select statements. It's a real good way to lose a lot of
>data fast!

While we're at it: never run ANY transformation program on the only
copy of the data, whether or not it deletes cases. Have the data in a
.SAV file (usually); load it; run the transformations; and save it
*under another name*. (I commonly use the convention that the name of
the new .SAV file is the same as the name of the .SPS syntax file that
created it.)

>I much prefer to create a flag using the conditions I plan to use to
>visually verify that I got it right. So, for example:
>
>compute flag = ~(sex = 'M' & hair_color = 'Brown').
>exe.
>
>Then visually confirm that the right people have been flagged for
>deletion. After you're convinced, change the compute statment to a
>select statement. But that's just me - I'm cautious.

A lot more people have died of too little caution than of too much.

But if I did this, I wouldn't even change the "compute" statement; I'd

SELECT IF flag EQ 1.

By the by, comparing CROSSTABS of (in this case) "sex" and "hair_color"
before and after selection, is another good test and confirmation.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Albert-Jan Roskam]

Hi,

In addition to Richard's advice: One trick to make it
impossible to loose your original data set is to put
its file attribute to read-only.

You can even do that from SPSS (v12 > (?)):
HOST COMMAND = ['attrib +r d:\temp\myfile.sav'].

Of course you can also do it in Windows, by just
ticking the right box under file properties.

Cheers!!
Albert-Jan


--- Richard Ristow <[hidden email]> wrote:

> At 06:22 PM 10/17/2007, Hal 9000 wrote:
>
>
> >select if ~(id = '123456').
> >
> >* (~ = "not").
> >
> >- or if you want to delete by some combination of
> classifications:
> >
> >select if ~(sex = 'M' & hair_color = 'Brown').
> >
> >*this deletes males with brown hair.
> >
> >Whatever you do, backup your original dataset
> before you start playing
> >around with select statements. It's a real good way
> to lose a lot of
> >data fast!
>
> While we're at it: never run ANY transformation
> program on the only
> copy of the data, whether or not it deletes cases.
> Have the data in a
> .SAV file (usually); load it; run the
> transformations; and save it
> *under another name*. (I commonly use the convention
> that the name of
> the new .SAV file is the same as the name of the
> .SPS syntax file that
> created it.)
>
> >I much prefer to create a flag using the conditions
> I plan to use to
> >visually verify that I got it right. So, for
> example:
> >
> >compute flag = ~(sex = 'M' & hair_color = 'Brown').
> >exe.
> >
> >Then visually confirm that the right people have
> been flagged for
> >deletion. After you're convinced, change the
> compute statment to a
> >select statement. But that's just me - I'm
> cautious.
>
> A lot more people have died of too little caution
> than of too much.
>
> But if I did this, I wouldn't even change the
> "compute" statement; I'd
>
> SELECT IF flag EQ 1.
>
> By the by, comparing CROSSTABS of (in this case)
> "sex" and "hair_color"
> before and after selection, is another good test and
> confirmation.
>
> =====================
> To manage your subscription to SPSSX-L, send a
> message to
> [hidden email] (not to SPSSX-L), with no
> body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send
> the command
> INFO REFCARD
>

Cheers!
Albert-Jan

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Did you know that 87.166253% of all statistics claim a precision of results that is not justified by the method employed? [HELMUT RICHTER]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Peck, Jon]

Actually, there is an entry on the SPSS File menu to mark a sav file as read only.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Albert-jan Roskam
Sent: Thursday, October 18, 2007 5:28 AM
To: [hidden email]
Subject: Re: [SPSSX-L] Deleting Cases

Hi,

In addition to Richard's advice: One trick to make it
impossible to loose your original data set is to put
its file attribute to read-only.

You can even do that from SPSS (v12 > (?)):
HOST COMMAND = ['attrib +r d:\temp\myfile.sav'].

Of course you can also do it in Windows, by just
ticking the right box under file properties.

Cheers!!
Albert-Jan

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Richard Ristow]

At 01:14 PM 10/18/2007, Peck, Jon wrote:

>Actually, there is an entry on the SPSS File menu to mark a sav file
>as read only.

So there is. It's implemented by issuing the syntax command
PERMISSIONS. New one on me: learn something new every day. (I wonder
when it came in? It is in 14; it doesn't seem to be in 9; and I don't
have any syntax manuals in between.)

CAUTION: The menu entry affects the file loaded as the current working
file. (If the current working file isn't associated with a .SAV file,
the menu entry is grayed out - reasonably enough.)

But if you want to save a file as read-only, save it and AFTERWARD use
the "Mark File Read-Only" entry (File menu). Here's a log, from trying
it. All the commands were clicked up from the menus and executed (not
pasted). The two PERMISSIONS commands were generated from the "Mark
File Read-Only" menu entry. The SAVE is from the "Save as..." menu
entry.

GET
   FILE='C:\Documents and Settings\Richard\My
Documents\Temporary\SPSS\Rubri'+
  'c data PRE- READ-ONLY.sav'.
DATASET NAME DataSet1 WINDOW=FRONT.

PERMISSIONS FILE = 'C:\Documents and Settings\Richard\My
Documents\Temporary\SPS
    S\Rubric data PRE- READ-ONLY.sav' /PERMISSIONS READONLY.

SAVE OUTFILE='C:\Documents and Settings\Richard\My
Documents\Temporary\SPSS\Rubr
    ic data SAVE AS READ-ONLY.sav'
  /COMPRESSED.

PERMISSIONS FILE = 'C:\Documents and Settings\Richard\My
Documents\Temporary\SPS
    S\Rubric data SAVE AS READ-ONLY.sav' /PERMISSIONS READONLY.

Interesting suggestion: on Save/Save AS dialog boxes, an option for
"save as read only", that would be implemented by emitting a
PERMISSIONS command following the SAVE.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Peck, Jon]

For the syntax lovers, there is PERMISSIONS subcommand on SAVE, so you can save and make read only in the same command.

 

The PERMISSIONS subcommand sets the operating system read/write permissions for the file.

READONLY

File permissions are set to read-only for all users. The file cannot be saved by using the same file name with subsequent changes unless the read/write permissions are changed in the operating system or the subsequent SAVE command specifies PERMISSIONS=WRITEABLE.

WRITEABLE

File permissions are set to allow writing for the file owner. If file permissions were set to read-only for other users, the file remains read-only for them.

 

 

Jon Peck

 

-----Original Message-----
From: Richard Ristow [mailto:[hidden email]]
Sent: Thursday, October 18, 2007 1:27 PM
To: Peck, Jon; [hidden email]
Subject: Re: Deleting Cases

 

At 01:14 PM 10/18/2007, Peck, Jon wrote:

 

>Actually, there is an entry on the SPSS File menu to mark a sav file

>as read only.

 

So there is. It's implemented by issuing the syntax command

PERMISSIONS. New one on me: learn something new every day. (I wonder

when it came in? It is in 14; it doesn't seem to be in 9; and I don't

have any syntax manuals in between.)

 

CAUTION: The menu entry affects the file loaded as the current working

file. (If the current working file isn't associated with a .SAV file,

the menu entry is grayed out - reasonably enough.)

 

But if you want to save a file as read-only, save it and AFTERWARD use

the "Mark File Read-Only" entry (File menu). Here's a log, from trying

it. All the commands were clicked up from the menus and executed (not

pasted). The two PERMISSIONS commands were generated from the "Mark

File Read-Only" menu entry. The SAVE is from the "Save as..." menu

entry.

 

GET

   FILE='C:\Documents and Settings\Richard\My

Documents\Temporary\SPSS\Rubri'+

  'c data PRE- READ-ONLY.sav'.

DATASET NAME DataSet1 WINDOW=FRONT.

 

PERMISSIONS FILE = 'C:\Documents and Settings\Richard\My

Documents\Temporary\SPS

    S\Rubric data PRE- READ-ONLY.sav' /PERMISSIONS READONLY.

 

SAVE OUTFILE='C:\Documents and Settings\Richard\My

Documents\Temporary\SPSS\Rubr

    ic data SAVE AS READ-ONLY.sav'

  /COMPRESSED.

 

PERMISSIONS FILE = 'C:\Documents and Settings\Richard\My

Documents\Temporary\SPS

    S\Rubric data SAVE AS READ-ONLY.sav' /PERMISSIONS READONLY.

 

Interesting suggestion: on Save/Save AS dialog boxes, an option for

"save as read only", that would be implemented by emitting a

PERMISSIONS command following the SAVE.