Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: Data conversion

From   Nick Cox <>
Subject   Re: st: Data conversion
Date   Wed, 23 May 2012 15:37:19 +0100

Would that it were true that -destring- should "work" for any string variable.

1. -destring- is designed to convert numbers presenting as strings to
the numbers they are entitled to be. Apart from some options to ignore
common extras such as commas or dollar signs, it doesn't really
purport to make numbers out of anything.

2. -destring- does not create its own coding system, by which I guess
you mean value labels. I think you are confusing -destring- with


I otherwise agree with Chelsea.  We need more details on what the data
look like and what you did.

I don't understand how you expect to convert "la", "lb", etc. to
numeric, but -encode- may be what you seek.


On Wed, May 23, 2012 at 3:23 PM, Chelsea Garneau
<> wrote:

> How did you convert strings to dates?  Were you using the destring
> command? This command should work for any string variable.  It will
> create it's own coding system, I believe alphabetically, but you can
> recode once it's in numerical format.

On Wed, May 23, 2012 at 9:49 AM, Hugo Fontan Köhler <> wrote:

>> I have a database with (originally in bDase format, convertd to xls in
>> Excel and imported in Stata) with a string variable in the format Ia,
>> Ib, and so on. I want to convert this variable in a numeric one. I
>> tried to generate a new variable with missing values (generate pZ=.)
>> and use replace (replace pZ = 1 if Z="1"), but I always get a type
>> mismatch error (r109). I managed to convert strings to dates but I
>> can't convert to numbers.

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index