Drop observations by a variable value if equal to one contained in a vector created with putmata or one contained in a local

[Valentino Santori]

Dear All,

I need to sample 40 observations from a dataset of 80 and then reload the full dataset to get the other 40 by dropping the ones in the first group.

import excel "...db.xlsx", sheet("...1") firstrow
generate n=_n
sample 50

Now I need to save the values of "n" variable of these sampled obs, re-import the dataset and drop the observations with n equal to this vector of values.
I thought about 2 ways of performing this task:
1 - save the vector of values in a matrix with putmata and re-importing the dataset

putmata group1=(n)
clear
import excel "...db.xlsx", sheet("...") firstrow
generate n=_n

but then I don't know how to drop the observations if n == one of the values in group1

2 - alternatively I thought I could save this vector of values in a local macro by using

levelsof n

after sampling, but I don't know how to store that output in a local.
I've done many attempts without success.

Thanks in advance for the support.

[Andrew Musau]

import excel "...db.xlsx", sheet("...1") firstrow clear
isid *
sample 50
tempfile selected
save `selected'
import excel "...db.xlsx", sheet("...1") firstrow clear
merge 1:1 * using `selected'
tab _merge

[Valentino Santori]

Hi Andrew Musau and thanks for your answer, save command return this error

. save `selected'
invalid file specification

EDIT: My bad, I was trying to exec one line at time.

[Andrew Musau]

You should run the code all at once in a do-file. You are probably running it in parts and thereby getting the error. But it should be more efficient to save the file to disk or use frames.

import excel "...db.xlsx", sheet("...1") firstrow clear
isid *
frame put *, into(match)
frame match: sample 50
frlink 1:1 *, frame(match)
replace match= !missing(match)
tab match

[Valentino Santori]

Thanks Andrew Musau I was editing the post but you were faster replying.
May I ask you if it is possible to do the thing I would like to in the ways I was trying ?

[Andrew Musau]

This is not tested.

import ...
gen n=_n
sample 50
levelsof n, local(levels) sep(,)

import...
gen n=_n
gen match= inlist(n, `levels')
tab match