On Wed, 30 Oct 2019, Marcin Błażejowski wrote:

On 30.10.2019 15:46, Sven Schreiber wrote:

Am 30.10.2019 um 10:24 schrieb Marcin Błażejowski:

Hi,

is it possible to change the ID number of given variable via script?

Hi Marcin, I don't think so (in general, without deleting other data),
because they are contiguous numbers used for indexing. Why do you want
to do that? Series names or named lists should typically be all you
need. (in theory)

It's tough to explain, but in short: we have database with houndrets of
series retrived from dbnomics, but after downloading we need to tune
some of them (interpolation, filtering and so on). We want to have a
tunned series next the original one, but all tunning staff we want to
run _after_ downloading the whole database. So, setting variable ID via
script would help us.

There's no command or function that specifically targets this operation, but it's fairly simple to re-order the series in a dataset via lists -- which can be treated as vectors (of series ID numbers) for manipulation -- plus the "store" and "open" commands.

Here's an example. Suppose we want to move the COLLEGE series from ID 9 to ID 2 in Ramanathan dataset 4-10:

<hansl>
open data4-10
list L = dataset
newpos = 2
oldpos = varnum(COLLEGE)
matrix mL = L
tmp = mL[newpos]
mL[newpos] = oldpos
mL[oldpos] = tmp
L = mL
store reordered.gdt L
open reordered.gdt
</hansl>

Allin

_______________________________________________
Gretl-users mailing list -- gretl-users@gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave@gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/

Thanks to all for help!!!

Marcin

-- 
Marcin Błażejowski