Short cut for repeat code, replacing just string value. Also

[Darcy Bornemann]

Two questions:

1. My main question is, Can I create a code short cut to replace the string variables within a set code.

I am analysing survey data. Questions where multiple responses are allowed, the survey system exports all the responses in one cell separated by a comma. I've split these in Stata using.

'split variable, p(,)'

Now to count/convert each instance I've started with this code.

generate variable2=1 if splitvarialble1=="stringvalueA" | splitvarialble2=="stringvalueA" | splitvarialble3=="stringvalueA" etc...
generate variable3=1 if splitvarialble1=="stringvalueB" | splitvarialble2=="stringvalueB" | splitvarialble3=="stringvalueB" etc...
generate variable4=1 if splitvarialble1=="stringvalueC" | splitvarialble2=="stringvalueC" | splitvarialble3=="stringvalueC" etc...

This has essentially counted the instances into a numeric value in a new column.

However, I have to replace the string variable in the code each time by hand. Is there a way to code this more simply or less copy past. I have a list of the string values (they are a fixed list from the survey).


Is it possible, or worth it, to arrange these up as a sequence within one cell eg. [0,1,1,1,0,0,0,1,0,1,1], where somewhere the string values are encoded to each position within the set (I hope this makes sense).

Many thanks,

Dr. DB

[Andrew Musau]

I cannot follow your post, and I suspect that others will have problems doing so as well. I recommend that you start with a data example (use dataex to present one), showing "this is what I have now" and "this is where I want to get to," rather than describing your attempted solution to the problem.

[Nick Cox]

I am embarrassed to report how long it took me to see that

gen anyA = inlist("A", var1, var2, var3)

is a way to do part of what I guess you want. I blame that on not just moderate stupidity but also the mathematics authors and teachers who by example taught me to write things like

Suppose x = 1.

Now they never told me to write

Suppose 1 = x

and why not -- because equalities are equalities and one is just as true as the other?

It's I think because = often (usually) means "Is" (*) and has asymmetric undertones, so there is a strong sense in which

x = 1

is intended to convey something like "our general thing is called x, and in this particular case its value is 1" and we don't write

1 = a

with reversed wording because that is not the way we think.

The tio can be found also at https://journals.sagepub.com/doi/pdf...867X0900900107 and https://journals.sagepub.com/doi/pdf...867X1101100308 but not in https://journals.sagepub.com/doi/pdf...867X0600600413

(*) This isn't original. Somewhere Donald Knuth quotes Nicolaas Govert de Bruijn to this effect.