Re: st: RE: From: D.Ronayne@warwick.ac.uk

[Eric Booth <ebooth@ppri.tamu.edu>]

<>


On Aug 13, 2010, at 3:16 PM, Martin Weiss wrote:
> 
> But the missings are back once the -merge-r is done. 

> Let`s summarize: We can twist it and bend it as much as we like, there are
> no "empty cells" in Stata...

Right.  From David's statement "without disturbing the other variables in the dataset"  I got the impression that he wanted to do what Excel does when you right click on some cells and choose to "Delete..">"Shift Cells Up".  Sorting moves all the rows together while extracting "x", removing/dropping the missing cells, and remerging does what I think David is asking for.

On Aug 13, 2010, at 3:24 PM, Maarten buis wrote:
> It is absolutely crucial that the values that
> belong to the same observations (i.e. a row in your
> dataset) always stay together. You can move entire 
> rows up or down, or you can move columns left or right,
> but you must never move individual cells up or down 
> or left or right. If that happens your entire dataset
> has instantly become completely worthless. 
> 

This is a good warning and is certainly true in (almost) all cases, but I assumed that David knows why he wants to move the data up the column and its effect on the data integrity.  
BTW, I have had situations where values in one or more columns are shifted down by a couple of rows, and I have to shift them up so that they are properly aligned (e.g., data that was extracted from a PDF or HTML page where the row spacing is interpreted by Acrobat or some other program).  

~ Eric

__
Eric A. Booth
Public Policy Research Institute
Texas A&M University
ebooth@ppri.tamu.edu
Office: +979.845.6754



On Aug 13, 2010, at 3:16 PM, Martin Weiss wrote:

> 
> <>
> 
> But the missings are back once the -merge-r is done. 
> 
> 
> 
> . li
> 
>     +---------------------------------+
>     | a   b       x            _merge |
>     |---------------------------------|
>  1. | 9   5    .998       matched (3) |
>  2. | 1   5    .658       matched (3) |
>  3. | 1   5   7.801       matched (3) |
>  4. | 1   5       .   master only (1) |
>  5. | 9   5       .   master only (1) |
>     |---------------------------------|
>  6. | 1   5       .   master only (1) |
>  7. | 9   5       .   master only (1) |
>  8. | 1   5       .   master only (1) |
> 
> 
> 
> Let`s summarize: We can twist it and bend it as much as we like, there are
> no "empty cells" in Stata...
> 
> 
> HTH
> Martin
> 
> 
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Eric Booth
> Sent: Freitag, 13. August 2010 22:12
> To: <statalist@hsphsun2.harvard.edu>
> Subject: Re: st: RE: From: D.Ronayne@warwick.ac.uk
> 
> If you want the non-missing cells to shift up to column and the other
> variables to stay fixed, you could create another dataset with just "x",
> remove the missings and merge it back in using a 1:1 merge on the row (_n):
> 
> ***********
> clear*
> 
> inp a x b 
> 9 0.998 5
> 1 . 5
> 1 . 5
> 1 . 5
> 9 0.658 5
> 1 . 5
> 9 7.801 5
> 1 . 5
> end
> 
> preserve
> keep x
> drop if mi(x)
> save using.dta , replace
> restore
> 
> drop x
> merge 1:1  _n using using.dta
> li
> ***********
> ~ Eric
> 
> __
> Eric A. Booth
> Public Policy Research Institute
> Texas A&M University
> ebooth@ppri.tamu.edu
> Office: +979.845.6754
> 
> 
> On Aug 13, 2010, at 3:00 PM, Martin Weiss wrote:
> 
>> 
>> <>
>> 
>> You can -sort- on the variable to assemble all the values other than
>> missings at the top. But the other variables will move as well...
>> 
>> ***********
>> clear*
>> 
>> inp x 
>> 0.998
>> .
>> .
>> .
>> 0.658
>> .
>> 7.801
>> .
>> end
>> 
>> gen byte othervar=_n
>> so x
>> 
>> list, noo
>> ***********
>> 
>> 
>> HTH
>> Martin
>> 
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu
>> [mailto:owner-statalist@hsphsun2.harvard.edu] 
>> Sent: Freitag, 13. August 2010 21:48
>> To: Stata List
>> Subject: st: From: D.Ronayne@warwick.ac.uk
>> 
>> Dear all,
>> 
>> I wish to delete "cells" (not observations, aka rows) that have missing
>> values for variable x only within x. Hence,
>> 
>> "drop if x==."
>> 
>> would not work because it will delete the observations across all
>> variables, to explain graphically, I want something that looks like this:
>> 
>> variable x
>> 0.998
>> .
>> .
>> .
>> 0.658
>> .
>> 7.801
>> .
>> 
>> to become:
>> 0.998
>> 0.658
>> 7.801
>> 
>> without disturbing the other variables in the dataset. Manually
>> manipulating in Excel is not an option because this is done half way
>> through a do file and must be done by written commands. Stata is very easy
>> to use to drop variables and observations, but can it delete specific
>> "cells"?
>> 
>> Can anyone help here?
>> 
>> Any advice much appreciated,
>> David


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/