Destringing values

[Martin Imelda Borg]

I have a column which is a mixture of numbers and words eg:

Less than 10
Null
More than 30

The above are stored as string values

TRY 1: - I understand this is used for numerical data as it won't work with encode
In order to destring them i have tried:

gen numerical = real (stringvalue)

the column came up with just hyphens - no blue text

then i tried:

TRY 2: - I understand encode should be used for just qualitative data
encode stringvalue, gen (numerical)

Some of them were successfully encoded, some weren't eg Less than 10 was successfully encoded
And I only managed to replace this value with a number 1 the rest I coudln't as they weren't encoded

What am I doing wrong?

[Bruce Weaver]

Hello Martin. What do you want each of those example strings to look like in the resulting numeric variable? Thanks for clarifying.

Less than 10 = ?
Null = ?
More than 30 = ?

[Martin Imelda Borg]

Hi Bruce, thanks for replying
I want:

1 = Less than 10
0 = Null
3 = More than 30


Hence I want to destring the values
Then replace with numbers

[Bruce Weaver]

Would you settle for whatever numeric values -encode- assigns? The following works for me using the 3 sample string values you gave:

Code:

clear
input str15 svar
"Less than 10"
"Null"
"More than 30"
end

encode svar, generate(nvar)
list
list, nolabel

Output:

. clear

. input str15 svar

                svar
  1. "Less than 10"
  2. "Null"
  3. "More than 30"
  4. end

.
. encode svar, generate(nvar)

. list

     +-----------------------------+
     |         svar           nvar |
     |-----------------------------|
  1. | Less than 10   Less than 10 |
  2. |         Null           Null |
  3. | More than 30   More than 30 |
     +-----------------------------+

. list, nolabel

     +---------------------+
     |         svar   nvar |
     |---------------------|
  1. | Less than 10      1 |
  2. |         Null      3 |
  3. | More than 30      2 |
     +---------------------+

.