String type mismatch error in frameappend

[Sonnen Blume]

Hi,

I am trying to append dataframes, and some of the variables are in diferent formats in different datasets. Its little difficult to cross-check which ones are different as there are too many variables, so I am wondering if there is any method to force the append despite the type mismatch.

. frameappend df1 df2 df3
namelist:  too many specified
r(103);

. frame dir
* default  1503 x 14162; hh_endline.dta
* df1      1387 x 7503; hh_baseline.dta
* df2      1359 x 8193; hh_midline.dta
* df3      1503 x 14162; hh_endline.dta

Note: frames marked with * contain unsaved data


. frameappend df2
shared variables in frames being combined must be both numeric or both string
type mismatch
r(109);

Please let me know.
Thanks in advance.

[Andrew Musau]

]so I am wondering if there is any method to force the append despite the type mismatch

No, you have to sort this out before appending.

[Andrew Musau]

If you use Stata's append command, it has a -force- option. So here's what you could do:

tempfile holding
frame change default
forval i=1/3{
    frame df`i': save `holding', replace
    append using `holding', force
}

But this may not help because the mismatched variables will contain missing values in observations corresponding to the appended dataset as described in the documentation:

force allows string variables to be appended to numeric variables and vice versa, resulting in missing values from the using dataset. If omitted, append issues an error message; if specified, append issues a warning message.

So that takes you back to #2.

[Clyde Schechter]

Its little difficult to cross-check which ones are different as there are too many variables,...

Mark Chatfield's -precombine-, available from SSC, will make this much easier for you.