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 on April 23, and its replacement, is already up and running.

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

Re: st: notes and labels and saving datasets

From   Louis Boakye-Yiadom <>
Subject   Re: st: notes and labels and saving datasets
Date   Thu, 2 Dec 2010 07:40:08 +0000 (GMT)

I think the issue Will raised is an important one, and I like Bill's reply. On the whole, however, I think it will be more useful for Stata to alert the user whenever a change has been made to the dataset, whether the change is in respect of a value, variable name, variable label, value label, sort order, variable order, notes, or value format. Interestingly, in Stata 10 (perhaps it's true for Stata 11 as well), the data-have-changed-since-last-saved alert is issued when an attempt is made to save data for which a variable has been renamed. However, the alert is not given in cases where a value label has been applied to a variable. Since Will's suggestion, if implemented, apparently will not cause any serious inconvenience, I strongly support it. Thanks.


--- On Wed, 1/12/10, William Gould, StataCorp LP <> wrote:

> From: William Gould, StataCorp LP <>
> Subject: Re: st: notes and labels and saving datasets
> To:
> Date: Wednesday, 1 December, 2010, 18:30
> Will Senn <>
> wrote, 
> > Is there a rationale for why Stata doesn't prompt to
> save files after 
> > notes or labels have been added/updated/deleted? If
> not, is this 
> > something that can be addressed in an update? What's
> the process of 
> > requesting such a thing?
> Yes, there is a rationale. The formats, labels, and notes
> that you add
> or change to the data are just dressing, not changes to the
> data
> themselves.  In the same vein, changing the sort order
> of the data is
> not considered a change.  The rule is that, if a
> change would not
> affect statistical results, if a change would affect only
> how results
> are displayed, then it does not count as a change.
> In terms of how to suggest improvements to Stata, you just
> did.
> Statacorp reads Statalist.  You could make the
> suggestion more formal
> by sending it to technical support, but that is not
> necessary.
> I admit that there have been times when I've made important
> changes to
> a dataset, where importance is measured in terms of my time
> to make
> the changes, and I've been irritated that Stata did not
> count that as
> a change when I forgot to save the dataset. 
> Generally, however, I
> like Stata's current behavior.  There have been times
> when the current
> behvaior has irritated me, but more often I use a dataset,
> generate
> results thinking that I have not changed the data in a
> meaningful way,
> and learn later that that I did.  In those situations,
> it's time for
> me to trace back and figure out what I did and whether the
> real change
> was intended.
> In this case, it would take a rising up of users -- such as
> on
> Statalist -- to convince us that we should change the
> data-have-changed-since-last-saved behavior.
> -- Bill
> *
> *   For searches and help try:
> *
> *
> *


*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index