Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: use13 : an experimental Stata command to import datasets in new Stata 13 format into older versions of Stata


From   Richard Williams <richardwilliams.ndu@gmail.com>
To   statalist@hsphsun2.harvard.edu, "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: use13 : an experimental Stata command to import datasets in new Stata 13 format into older versions of Stata
Date   Wed, 19 Jun 2013 17:45:47 -0400

At 05:31 PM 6/19/2013, Sergiy Radyakin wrote:
On Wed, Jun 19, 2013 at 5:19 PM, Richard Williams
<richardwilliams.ndu@gmail.com> wrote:
> I tried converting this file with Stat/Transfer 12. It ran but the variable

Richard, which file? What was the original variable type and storage type?

It was the toy 2 observation file Robert created that included the command

gen s = "x" * 5000

Complete code is below in the earlier messages.

So, S/T seems to be following Nick's (that he did not recommend) strategy of truncating the variable. Stata 13 and -use13- either drop the variable or refuse to do anything at all. I think I prefer the latter -- I'd like to be explicitly told how the conversion is failing rather than be expected to have read and memorized the help.


> was saved in str244 format and there was no warning that this was happening.

For strL variables -use13- reports as follows:

*   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *
 *   *   *
Converting file ...{skipped}....\use13\strltest.dta
Variable xstrl has type strL which is not supported in this version of
Stata. This variable is dropped.
*   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *   *
 *   *   *
Sergiy Radyakin



> The S/T help says
>
> "Stata Version 13 includes a  new strl data type.  These are potentially
> very long, variable length strings.   Stat/Transfer will write strings
> longer than a threshold value to Stata strl's rather than strings.  The
> threshold, by default, is 32 characters, but you can change this on the
> Options (1) screen.  Strings longer than  Stata's width limit  of 244
> characters will always be written to strls.   Dates are written to Stata's
> internal date format."
>
> I think the moral is if you are going to have monster strings, you better
> have everybody working with Stata 13.
>
> I suspect S/T just assumes you know what you are getting when you ask for a
> conversion, i.e. you have to read the help because it won't otherwise warn
> you what compromises had to be made.
>
>
> At 03:09 PM 6/19/2013, Robert Picard wrote:
>>
>> The -saveold- command of Stata 13 returns an error if the data
>> contains a long string:
>>
>> . clear
>>
>> . set obs 2
>> obs was 0, now 2
>>
>> . gen s = "x" * 5000
>>
>> . saveold test.dta
>> data cannot be saved in old format
>>     Data contain strL or str#, #>244, and prior releases of Stata
>> would not know how to process these variables.  Either drop the
>>     variables or use recast with the force option to change them to
>> str244.
>> r(459);
>>
>>
>>
>> On Wed, Jun 19, 2013 at 2:25 PM, Sergiy Radyakin <serjradyakin@gmail.com>
>> wrote:
>> > On Wed, Jun 19, 2013 at 1:11 PM, Nick Cox <njcoxstata@gmail.com> wrote:
>> >> I have no access to Stata 13, but there is precisely no point to
>> >> -saveold- if Stata 12 is presented with data features it can not
>> >> handle. So, at best long string variables will get truncated at 244
>> >> characters. That's my guess.
>> >
>> > The idea is to be creative and make the content accessible in Stata 12.
>> > There
>> > are several strategies: splitting long variables into chunks,
>> >  (resume-->resume_1, resume_2, resume_3....)
>> > each no longer than 244, or dumping the long variables into a text file,
>> > from which the users can pull the strings using alternative methods,
>> > etc.
>> > The simplest approach is to truncate the data, but it is not
>> > necessarily the best.
>> >
>> > But the bulk of the users are not going to use the new features of Stata
>> > 13
>> > right away. Just as many of the datasets that are shipped with Stata 12
>> > can
>> > be resaved for Stata 9 without any losses. Consider auto.dta for
>> > example.
>> > Secondly, since the features of Stata 13 were kept out of public view,
>> > noone
>> > was planning for them in advance, and having the possibility of long
>> > strings
>> > would not change the questionnaires being designed this month, and the
>> > data
>> > collected using them in the next half a year.
>> >
>> > The feature that lacks in Stata is really: savemin (imaginary name),
>> > which would
>> > examine the file and save it in the lowest possible version that would
>> > not incur
>> > losses (that is really a job for save, but let's not get into that).
>> > With that, many
>> > datasets will end up somewhere in the v4-v7 range (imho).
>> >
>> > Another strategy could have been dual compatibility: SPSS v7 can read
>> > SPSS
>> > v11 datasets. It does not understand everything, but the core of data
>> > can be read,
>> > and decoration is often (not always, but often) not essential.
>> >
>> >>
>> >> I hope you don't get troubled by users who imagine that you have found
>> >> a cheap way for them to upgrade without paying!
>> >
>> > Nothing in the announcement implies that. It is not an upgrade for Stata
>> > 12 that
>> > delivers features of Stata 13. It is a way to read data from Stata 13
>> > subject to
>> > a) limitations of the existing Stata versions and b) my own goals and
>> > time/effort
>> > constraints. Not sure whether the same argument applies to e.g.
>> > StatTransfer,
>> > which is doing a similar conversion, though having had more time for
>> > implementation (and [hypothetically] more insider information).
>> >
>> > Best, Sergiy
>> >
>> >
>> >> Nick
>> >> njcoxstata@gmail.com
>> >>
>> >>
>> >> On 19 June 2013 18:05, Sergiy Radyakin <serjradyakin@gmail.com> wrote:
>> >>> And just as long they continue to "forget" doing this.
>> >>>
>> >>> Also, I am not sure what is the behavior of -saveold- in Stata 13.
>> >>> What does it do with the long strings? The documentation is only
>> >>> saying "saves the dataset currently in memory on disk ... in Stata 12
>> >>> format". How is conversion done?
>> >>>
>> >>> Sergiy.
>> >>>
>> >>> On Wed, Jun 19, 2013 at 1:00 PM, Nick Cox <njcoxstata@gmail.com>
>> >>> wrote:
>> >>>> Note that getting your colleagues using Stata 13 to use -saveold- to
>> >>>> save their datasets is the age-old solution to this.
>> >>>>
>> >>>> Nick
>> >>>> njcoxstata@gmail.com
>> >>>>
>> >>>>
>> >>>> On 19 June 2013 17:54, Sergiy Radyakin <serjradyakin@gmail.com>
>> >>>> wrote:
>> >>>>> -use13-  is an experimental Stata command to import datasets in new
>> >>>>> Stata 13 format into older versions of Stata
>> >>>>>
>> >>>>> The dataset format has changed in the new version of Stata to be
>> >>>>> released on June 24, 2013 to accommodate new features.
>> >>>>>
>> >>>>> The command will help users who can't upgrade immediately, but need
>> >>>>> to
>> >>>>> load the data produced by Stata 13 (e.g. by their colleagues who
>> >>>>> have
>> >>>>> upgraded) during the transition period.
>> >>>>>
>> >>>>> The use, limitations, and installation instructions are described
>> >>>>> here:
>> >>>>> http://www.adeptanalytics.org/radyakin/stata/use13/use13.htm
>> >>>>>
>> >>>>> Best, Sergiy Radyakin
>> >>>>> *
>> >>>>> *   For searches and help try:
>> >>>>> *   http://www.stata.com/help.cgi?search
>> >>>>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> >>>>> *   http://www.ats.ucla.edu/stat/stata/
>> >>>> *
>> >>>> *   For searches and help try:
>> >>>> *   http://www.stata.com/help.cgi?search
>> >>>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> >>>> *   http://www.ats.ucla.edu/stat/stata/
>> >>> *
>> >>> *   For searches and help try:
>> >>> *   http://www.stata.com/help.cgi?search
>> >>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> >>> *   http://www.ats.ucla.edu/stat/stata/
>> >> *
>> >> *   For searches and help try:
>> >> *   http://www.stata.com/help.cgi?search
>> >> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> >> *   http://www.ats.ucla.edu/stat/stata/
>> > *
>> > *   For searches and help try:
>> > *   http://www.stata.com/help.cgi?search
>> > *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> > *   http://www.ats.ucla.edu/stat/stata/
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/
>
>
> -------------------------------------------
> Richard Williams, Notre Dame Dept of Sociology
> OFFICE: (574)631-6668, (574)631-6463
> HOME:   (574)289-5227
> EMAIL:  Richard.A.Williams.5@ND.Edu
> WWW:    http://www.nd.edu/~rwilliam
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam

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


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index