Re: st: reading data with missing obs

[Michael Hanson <mshanson@mac.com>]

On Feb 8, 2009, at 3:14 AM, Glen Waddell wrote:

> I am trying to read in data from multiple xml files (although this  
> issue is not specific to this format, I believe) in which some of  
> the sheets within some of the files have variable names in the  
> first row and nothing else.  That is, all variables are missing in  
> some of the sheets within some of the files.
>
> In reading in these files, Stata only appears to read in the first  
> variable... not continuing to subsequent columns.  If I had but a  
> few files I would brute force this by filling the empty cells with  
> some unique character, read them in and then drop the obs  
> accordingly.  However, I have many of these sheets to be read in,  
> so automating it would be quite valuable.


I recently had a similar situation in which columns of data (from an  
Excel spreadsheet saved in xml format) started with missing values,  
and thus were not read in correctly with the basic -xmluse- command.   
After some experimentation, I found I had to read the data as  
"allstring" (an option of -xmluse-) and then -destring- the desired  
series after Stata imported them as string variables in order to  
missing values to be captured correctly.  For example:

xmluse [file.xml], doctype(excel) cells([b2:z300]) allstring firstrow  
missing nocompress
destring [varlist], ignore([stray_text]) replace

Note that you will need to replace the items in [square brackets]  
with those appropriate to your problem.  Also note that this solution  
is specific to xml formatted data, even if the problem (as you  
suggest) is not.

Hope this helps,
Mike

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/