[Abiye Kalaiwo]

You just need the one variable.

dataex SystolicBP if missing(real(SystolicBP))

dataex SystolicBP if missing(real(SystolicBP))

----------------------- copy starting from the next line -----------------------

* Example generated by -dataex-. To install: ssc install dataex
clear
input str4 SystolicBP
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
""
end

------------------ copy up to and including the previous line ------------------

Listed 100 out of 53146 observations
Use the count() option to list more


If what you are trying to achieve is to filter out the entries that have the characters, there is a way I get that. Let me know if that is what you are driving at.

[Andrew Musau]

If what you are trying to achieve is to filter out the entries that have the characters, there is a way I get that. Let me know if that is what you are driving at.

Yes, that is what is needed. You can keep a couple and then use dataex to extract them and post here.

[Abiye Kalaiwo]

Yes, that is what is needed. You can keep a couple and then use dataex to extract them and post here.

list SystolicBP if !missing(SystolicBP) & missing(real(SystolicBP)), noobs

+----------+
| Systol~P |
|----------|
| 120` |
| 130- |
| `100 |
| 120- |
| 100` |
|----------|
| 10-0 |
| `130 |
| - |
| `120 |
| 101` |
|----------|
| `90 |
| `117 |
+----------+



. dataex SystolicBP if !missing(SystolicBP) & missing(real(SystolicBP))

----------------------- copy starting from the next line -----------------------

* Example generated by -dataex-. To install: ssc install dataex
clear
input str4 SystolicBP
"120`"
"130-"
"`100"
"120-"
"100`"
"10-0"
"`130"
"-"   
"`120"
"101`"
"`90" 
"`117"
end

------------------ copy up to and including the previous line ------------------

Listed 12 out of 53146 observations

[Andrew Musau]

The grave accent is used in local macros, so that is why destring has problems if you include it within -ignore-. I will look into this, but the following should work:

* Example generated by -dataex-. To install: ssc install dataex
clear
input str4 SystolicBP
"120`"
"130-"
"`100"
"120-"
"100`"
"10-0"
"`130"
"-"  
"`120"
"101`"
"`90"
"`117"
end

gen wanted= real(ustrregexra(SystolicBP, "[^\d]+", ""))

Res.:

. l, sep(0)

     +-------------------+
     | Systol~P   wanted |
     |-------------------|
  1. |     120`      120 |
  2. |     130-      130 |
  3. |     `100      100 |
  4. |     120-      120 |
  5. |     100`      100 |
  6. |     10-0      100 |
  7. |     `130      130 |
  8. |        -        . |
  9. |     `120      120 |
 10. |     101`      101 |
 11. |      `90       90 |
 12. |     `117      117 |
     +-------------------+

[Abiye Kalaiwo]

Thanks for this. Unfortunately it is a bit confusing for me as I am pretty new to Stata. Would appreciate some breakdown.

[Andrew Musau]

gen wanted= real(ustrregexra(SystolicBP, "[^\d]+", ""))

The -real- function converts a string to a numeric variable with values either being numbers or missing (only number strings can be converted to real numbers). See

help real()

The rest is just deleting all nonnumeric characters in your variable and then passing the result through the -real- function. See a combination of the ICU Regular Expressions Manual and the -ustrregexra- function to gain a deeper understanding.

help ustrregexra()

[Abiye Kalaiwo]

So I have another string related question and I'm wondering if it is ok to post here or start another post? I do not really know what the rules are around that. Someone please help me. Thanks

[William Lisowski]

It would be best to start a new post, with a title that describes your new question. Members often decide whether or not they are going to follow a topic, and members who have decided not to follow this topic may not see your new question if you post it here.

[Abiye Kalaiwo]

It would be best to start a new post, with a title that describes your new question. Members often decide whether or not they are going to follow a topic, and members who have decided not to follow this topic may not see your new question if you post it here.

Thanks. This is helpful. I will start a new post.