Liebe FDZlerInnen,
gibt es vll. irgendwo eine Tabelle/Übersicht, der man unmittelbar entnehmen kann, welche Variablen in welchen Wellen erhoben wurden und in welchen nicht (z.B. für die SC3)? Z.B. soetwas:
Welle 1 Welle 2 Welle 3 ....
Variable_1 X NA X
Variable_2 X X X
Variable_3 X NA X
....
Wäre super hilfreich!
LG
Nicolas
Hallo Nicolas!
Ich habe hier ein kleines do-file, damit kann man für jeden SUF passende Excel-Sheets ganz fix erzeugen. Du musst einfach nur die locals anpassen und schon kann es losgehen.
Viele Grüße
Dietmar
version 12
// define locals
quietly {
noisily: display as text "specify locals before your start!!!"
local cohort SC3
local version 11-0-1
local sufpath <PATH_TO_DATA> // e.g. C:/Users/JaneDoe/Desktop/Data/SC3_11-0-1/Stata14
local level D
local targetfile <PATH_TO_EXCELSHEET>/<NAME_OF_EXCELSHEET> // e.g. `"C:/Users/JaneDoe/Desktop/NEPS_Stuff/SC3_varnames_X_wave.xlsx"'
noisily: display as text "install {bf:nepsmgmt} to use this syntax!!!"
net install nepsmgmt, from(http://nocrypt.neps-data.de/stata)
noisily: display as text "find files to create varnames x wave - list"
filesearch .*\.dta, dir("`sufpath'") regex local(suffiles)
foreach suffile of local suffiles {
if regexm("`suffile'","^`cohort'_(.*)_`level'_`version'\.dta$") local shortname = regexs(1)
if !regexm("`shortname'","^sp") { // skip episode files, because wave doesn't make sense
capture: describe wave using "`sufpath'/`suffile'" // check if wave is there
if _rc == 0 {
local appendlist ""
noisily: display as text "{bf: processing `shortname'}... found variable {it:wave} opening dataset!!!"
use "`sufpath'/`suffile'" , clear
preserve
noisily: display as text "{tab}>> create table of variable names and positions to sort lines later"
describe, replace
keep position name
tempfile poslist
save "`poslist'", replace
restore
noisily: display as text "{tab}>> slicing data into separate waves..."
sum wave
forvalues wave=-29/`r(max)' {
preserve
keep if wave == `wave'
count
local old_obs `r(N)'
if `old_obs' > 0 {
noisily: display as text "{tab}{bf:processing wave `wave'}... keeping wave-`wave'-data"
noisily: display as text "{tab}{tab}drop all variables only filled with design missings"
foreach var of varlist _all {
if regexm("`: type `var''","str") count if regexm(`var',"^-54")
else count if `var' == -54
if `r(N)'==`old_obs' drop `var'
}
describe, replace clear
keep name varlab
if `wave' > 0 generate wave_`wave' = 1
else{
local wave=(`wave')*(-1)
generate wave_min`wave' = 1
}
noisily: display as text "{tab}{tab}save list of filled variables"
tempfile varlist_wave`wave'
save "`varlist_wave`wave''", replace
local appendlist : list appendlist | varlist_wave`wave'
}
restore
}
local maxappend = `: word count `appendlist''
noisily: display as text "{tab}{bf:merging wave-specific file lists} to create varnames-X-wave-matrix"
if `: word count `appendlist'' >= 1 {
clear
local cntr 0
foreach file of local appendlist {
local `++cntr'
if `cntr' == 1 use "`file'", clear
else merge 1:1 name using "`file'", nogenerate noreport nolabel
foreach var of varlist wave_* {
replace `var' = 0 if missing(`var')
}
if `cntr' == `maxappend' {
merge 1:1 name using "`poslist'", nogenerate noreport nolabel
sort position
drop position
export excel using `"`targetfile'"', firstrow(variables) sheet("`shortname'")
}
}
noisily: display as text "{tab}{tab}{bf:export list} to {it:`targetfile'} - sheet: `shortname'"
}
noisily: display as text "{dup 60:+}
}
}
}
noisily: display as text "{bf:...done!!!}"
}
clear
exit 0
end
// EOF
Merci Dietmar,
probiere ich aus und melde mich nochmal, falls etwas nicht klappt.
LG
Nicolas