www.cmm.bristol.ac.uk/forum

Hi,

I previously have had no trouble opening STATA (.dta) datasets in MLwiN and
saving as a worksheet in MLwiN.

For awhile now, however, whenever I do that in MLwiN the string variables
are not "read" and take on a constant value. Moreover, when I try to write
a model using this not-well-read dataset, MLwiN is VERY SLOW and crashes.

But it works fine when I go back to my old worksheets in which the STATA
.dta file was read in fine it works fine.

For example, I can no longer read in even the same .dta sets that I have created MLwiN worksheets from without having the above problems

I do not know why I cannot read in .dta files anymore and cannot figure out what the problem is and want to be able to work with
STATA datasheets in MLwIN as soon as possible.

Thanks for your help.

There is unfortunately a bug in version 2.18 which means that it does not correctly import columns defined as text (this affects Stata, SPSS and pasted ASCII text). This will be fixed in the next release.

Possible workarounds are to convert the column into numbers with labels before importing, to use an older version of MLwiN to import the data, or to clear the generated labels via "toggle categorical" and then re-enter them manually or via the copy/paste category functionality.

Hi, MLwiN has been upgraded today and this bug is now fixed