String to numeric

[David Page]

Sorry - I know this is very basic, I'm a relatively novice.

If I want to manipulate a variable that is a word into a number, how do I do it? I've been able to answer most of my Stats questions using online tutorials thus far, but I haven't been able to successfully do this.

As an example, let's say I have a variable called letter. I have two options for the variable letter, "alpha" and "beta." I want to change alpha to 1 and beta to 2. I think it's some variation of destring, but I can't get the syntax right. I've had to do this in the past with small data sets and simply manually manipulated the data sheet, but I have larger data set for which manual manipulation would be cumbersome.

Thanks!

[Chen Samulsion]

You'd better provide data example using dataex command. Without data we have to guess your data structure and variable format. As you described, letter is a string variable which has two values, alpha and beta. Your aim is transforming string variable to numeric variable. So you can just type in command window

encode letter, gen(new_letter)

And following codes allow you to check your transforming result:

list letter new_letter in 1/5
list letter new_letter in 1/5, nolabel

[Nick Cox]

https://www.stata-journal.com/articl...article=dm0098 offers a tutorial review of this territory. It benefits from puzzled questions here on Statalist.

[Marcos Almeida]

Just try typing “Stata string to numeric” in the web.

[David Page]

Just try typing “Stata string to numeric” in the web.

I tried that but couldn’t get it to work.

[David Page]

You'd better provide data example using dataex command. Without data we have to guess your data structure and variable format. As you described, letter is a string variable which has two values, alpha and beta. Your aim is transforming string variable to numeric variable. So you can just type in command window

encode letter, gen(new_letter)

And following codes allow you to check your transforming result:

list letter new_letter in 1/5
list letter new_letter in 1/5, nolabel

Thanks! I’ll give that a try this afternoon. I’m trying to figure out how to do things like change male/female variables to 1/2 or white/black/Hispanic to 1/2/3.

What exactly does the 1/5 that you’re writing mean? Does that isolate the first letter?

Again, sorry - I’m kind of a novice. I’ve found the online Stata tutorials very helpful, but I get lost in some of them.

[Chen Samulsion]

1/5 is range which means list first five observations. I recommend -list- command just convenience you to check whether the encode result is what you desire. Please take care of the nolabel option.

[Nick Cox]

On-line tutorials are widely appreciated, but your first resource always is, or should be, the help in Stata.

Given a puzzle about list, then

help list

and if you follow the link to the Quick Start, the use of in is explained immediately.