Extracting specific information from rows

[SarahW13]

I have a dataset in which each row is a long block of text, and I need to extract specific information from each row.

Below is an example:

id       Text

1        Physical Exam. Vital Signs. BP: 134/93

2        Patient's physical exam was notable for BP of 142/100

3        Physical Exam. BP: 100/80

Below is what I need:

id       Text

1        BP: 134/93

2        BP of 142/100

3        BP: 100/80

Any advice?

[PeterClemmensen]

Do something like this

data have;
input id Text :$200.;
infile datalines4 dlm=',';
datalines;
1,Physical Exam. Vital Signs. BP: 134/93
2,Patient's physical exam was notable for BP of 142/100
3,Physical Exam. BP: 100/80
;

data want;
    set have;
    NewString=substr(Text, index(Text, 'BP'));
run;

[alexpat]

data cardiac;
input id Text :$200.;
infile datalines4 dlm=',';
datalines;
1,Physical Exam. Vital Signs. BP: 134/93
2,Patient's physical exam was notable for BP of 142/100
3,Physical Exam. BP: 100/80
;

data strepto;
set cardiac;
NewString=substr(Text, index(Text, 'BP'));
run;

[Jagadishkatam]

Alternatively with prxchange function

data have;
input id Text&$100.;
text=prxchange('s/(.*)(bp.*)/$2/i',-1,text);
datalines4;
1 Physical Exam. Vital Signs. BP: 134/93
2 Patient's physical exam was notable for BP of 142/100
3 Physical Exam. BP: 100/80
;;;;

[alexpat]

Rows 1-3

Total rows: 3Total columns: 3

idTextNewString

1

1

Physical Exam. Vital Signs. BP: 134/93

BP: 134/93

2

2

Patient's physical exam was notable for BP of 142/100

BP of 142/100

3

3

Physical Exam. BP: 100/80

BP: 100/80

[Ksharp]

data have;
input id Text :$200.;
infile datalines4 dlm=',';
datalines;
1,Physical Exam. Vital Signs. BP: 134/93
2,Patient's physical exam was notable for BP of 142/100
3,Physical Exam. BP: 100/80
;

data want;
    set have;
    p=prxmatch('/\bBP\b/i',text);
    if p then NewString=substr(Text, p);
run;

proc print;run;

[alexpat]

data sle;
set autoimmun;
p=prxmatch('/\bBP\b/i',text);
if p then NewString=substr(Text, p);
run;

Thanks Ksharp, another way to estract ///

[ballardw]

---

@SarahW13 wrote:

I have a dataset in which each row is a long block of text, and I need to extract specific information from each row.

Below is an example:

id       Text

1        Physical Exam. Vital Signs. BP: 134/93

2        Patient's physical exam was notable for BP of 142/100

3        Physical Exam. BP: 100/80

Below is what I need:

id       Text

1        BP: 134/93

2        BP of 142/100

3        BP: 100/80

Any advice?

---

is BP always the last item entered? Is it always recorded with the / dividing the measurements?