topic Re: How to read from a text in SAS Programming

How to read from a text

JAR — Fri, 21 Apr 2017 04:24:45 GMT

I am working on a project that will help children learn faster/better. I have got a raw data file with some dictionary data. The follwong have and want may be self explanatory:
Mostly the data is created systematically. Here are some characteristics I have found which may help in programming:

Word: Begin with no space
Pronounciation: in "(" and ")"

Etymology: in "[" and "]"

If pronounciation and etymology is given, the word is repeated for defintion
Defintion: start with a -space- character

Examples: starts with a year

Can someone please help? (Please see the screenshot for clarity)

Jijil Ramakrishnan

Re: How to read from a text

Reeza — Fri, 21 Apr 2017 04:15:36 GMT

What is the question here?

Re: How to read from a text

JAR — Fri, 21 Apr 2017 04:16:31 GMT

Any suggestion to program this?

Re: How to read from a text

Reeza — Fri, 21 Apr 2017 04:18:42 GMT

Program what? You haven't explained what the question is or what you're trying to do. We can't read minds.

Re: How to read from a text

Reeza — Fri, 21 Apr 2017 04:20:22 GMT

My bad. You wrote on the images. That wasn't clear at first.

Thats going to need regular expressions which is not my forte.

Re: How to read from a text

Ksharp — Fri, 21 Apr 2017 14:44:31 GMT

That is really uneasy.

data x;
infile '/folders/myfolders/samplerawfile.txt' lrecl=32767 length=len;
input x $varying32767. len;
if missing(x) then delete;
run;
data temp;
 merge x x(firstobs=2 rename=(x=_x));
 if prxmatch('/^\(.+\)\[.+\]$/',strip(_x)) then group+1;
 drop _x;
run;

data temp;
 set temp;
 by group;
 length name temp $ 80;
 retain temp;
 if first.group then do; n=0;call missing(temp);end;
 n+1;
 if n=1 then do;name='word';temp=x;end;
  else if prxmatch('/^\(.+\)\[.+\]$/',strip(x)) then name='pronounce';
   else if prxmatch('/^[\w\p]+:/',strip(x)) then name='example';
    else name='meaning';
 if temp=x and n ne 1 then delete;
 drop temp n ;
run;
data temp;
 set temp;
 length _name $ 80;
 by group;
 if first.group then do;idx=0;n=0;end;
 if name='meaning' then do;
   idx+1;
   _name=catx('_',name,idx);
 end;
 drop n;
run;
data temp;
 set temp;
 by group idx notsorted;
 if first.idx then m=-1;
 m+1;
 if name='example' then _name=cats(cats('meaning',idx),'_',cats(name,m));
  else if missing(_name) then _name=name;
run;

proc transpose data=temp out=want;
by group;
id _name;
var x;
run;
proc print data=want(obs=50) noobs;run;

Re: How to read from a text

JAR — Fri, 21 Apr 2017 18:12:42 GMT

I thank you very much, Mr. KSharp! You have done more than what I had requested. Is there any study indicating a correlation between knowledge and magnanimity? Thank you very much!

Re: How to read from a text

Ksharp — Sat, 22 Apr 2017 03:03:37 GMT

I can not guarantee my code can get what you want.

You need know what kind of rules you should take into account.

For example.

prxmatch('/^[\w\p]+/')

also could change into

prxmatch('/^\w+/')

it is depended on your data.

" Is there any study indicating a correlation between knowledge and magnanimity?"

You mean make some statistical model ?

Plz post it at Statisical forum .

Re: How to read from a text

Reeza — Sat, 22 Apr 2017 03:08:25 GMT

" Is there any study indicating a correlation between knowledge and magnanimity?"

You mean make some statistical model ?

Plz post it at Statisical forum .

@Ksharp I believe that was a complement of sorts, not a question. Smarter people are helpful 🙂

Re: How to read from a text

Ksharp — Sat, 22 Apr 2017 03:15:58 GMT

@Reeza

Actually , I don't know what OP means.

I am considering about some category data analysis, such as contingency table analysis.

Re: How to read from a text

Ksharp — Sat, 22 Apr 2017 03:19:48 GMT

@Reeza

You mean it is a compliment ?

Re: How to read from a text

Reeza — Sat, 22 Apr 2017 03:52:58 GMT

@Ksharp wrote:

@Reeza

You mean it is a compliment ?

Yeah it's a compliment. Sorry, I misused the incorrect but similar sounding word. English is a great language sometimes.