Am Montag, den 14.09.2009, 06:27 +0200 schrieb Si Yue Li:
> as a beginner, I am having some trouble importing large excel datatables
> containing datastream output into stata long format. I checked with
> faq/help files and think that the reshape command combined with some
> data manipulations must be the answer. Right now I can get the data to
> look like this in Stata:
> |v1|v2|v3|v4|v5|...
> |.|ID code1|ID code2|ID code3|ID code4|...
> |date1| Xij|Xij|Xij|Xij|...
> |date2| Xij|Xij|Xij|Xij|...
> |...|...|...|...|...
>
> where Xij = Data
>
> However from my past Stata experiences the usual long form is much
> better to work with:
>
> |ID code1|date1|Xij|
> |ID code1|date2|Xij|
> |...|...|...|
> |ID code2|date1|Xij|
> |ID code2|date2|Xij|
> |...|...|...|
>
> I think reshape will solve this for me, but I couldn't get anything to
> work...
Which command do you have used to read in your Excel file? I have
created the file foo.raw as tab separated values file. You get this
format by saving as "plain text (.tsv)" in Excel. The Stata command
. cat foo.raw, showtabs
should produce something like this:
.<T>ID code1<T>ID code2<T>ID code3<T>ID code4
date1<T> 15<T>15<T>15<T>15
date2<T> 15<T>15<T>15<T>15
whereby <T> indicates a tab. If I read in foo.raw with
. insheet using foo.raw
I get
. l, clean
v1 idcode1 idcode2 idcode3 idcode4
1. date1 15 15 15 15
2. date2 15 15 15 15
which could be reshaped to long as follows
. reshape long idcode, i(v1)
Many regards
Uli
*
* 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/
