spset after merge: spatial details disappear

[Dario Maimone Ansaldo Patti]

Dear All,

I posted some time ago a similar post. I thought to have solved my problem, but it is still there. I use stata15. Basically, I convert a shapefile into a stata one using spshape2dta. It works fine:

 spshape2dta shapecountries
  (importing .shp file)
  (importing .dbf file)
  (creating _ID spatial-unit id)
  (creating _CX coordinate)
  (creating _CY coordinate)

  file shapecountries_shp.dta created
  file shapecountries.dta     created

If I spset the file shapecountries.dta I get:

spset
  Sp dataset shapecountries.dta
    data:  cross sectional
    spatial-unit id:  _ID
    coordinates:  _CX, _CY (planar)
    linked shapefile:  shapecountries_shp.dta

Then I try to merge this file using another one:

merge m:m NAME using vp1

    Result                           # of obs.
    -----------------------------------------
    not matched                  3
    from master                   0  (_merge==1)
    from using                     3  (_merge==2)

    matched                       248  (_merge==3)
    -----------------------------------------

Finally, I check that everything is fine with the spset (as far as I read, although I modify the file somehow - as merging it with another one - it should preserve the previous spset):

spset
invalid panel setting
Type xtset to view panel data settings.

Why this is the case? Do you have any suggestion? Just for your reference, I upload the files used above. Thanks in advance.

Dario

https://dl.dropboxusercontent.com/u/27015126/data.zip

[Kevin Crow (StataCorp)]

It looks like your vp1 dataset has a bad xtset setting. Type xtset, clear and then resave the dataset. Also, it looks like your dataset has duplicate NAME values which will cause spset to error.

. use vp1, clear

. xtset
variable year not found
r(111);

. xtset, clear

. save, replace
file vp1.dta saved

. duplicates list NAME

Duplicates in terms of NAME

  +----------------------------+
  | group:   obs:         NAME |
  |----------------------------|
  |      1    121   Kyrgyzstan |
  |      1    122   Kyrgyzstan |
  |      2    202     Slovakia |
  |      2    203     Slovakia |
  +----------------------------+

. merge m:m NAME using shapecountries

    Result                           # of obs.
    -----------------------------------------
    not matched                             3
        from master                         3  (_merge==1)
        from using                          0  (_merge==2)

    matched                               248  (_merge==3)
    -----------------------------------------

. drop if _merge == 1
(3 observations deleted)

. spset
  Sp dataset vp1.dta
                data:  cross sectional
     spatial-unit id:  _ID
         coordinates:  _CX, _CY (planar)
    linked shapefile:  shapecountries_shp.dta
variable _ID does not uniquely identify the observations
r(459);

. list NAME _ID if _ID == 105

     +----------------+
     |     NAME   _ID |
     |----------------|
200. | Slovakia   105 |
201. | Slovakia   105 |
     +----------------+

[Dario Maimone Ansaldo Patti]

Dear Kevin,

thanks for your reply. Actually, my main problem was with spset. I knew NAME may not uniquely identify observations. The file vp1 comes from collapsing a panel data. Apparently, the file still contained the previous xtset.

Thank you very much for your help.

Best,

Dario