cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
KachiM
2 weeks ago
  • 585 Posts
  • 13 Likes Given
  • 54 Solutions
  • 198 Likes Received

Re: Need an efficient way to combine datasets?

by Rhodochrosite | Level 12 in SAS Programming
‎01-28-2015 11:40 AM
‎01-28-2015 11:40 AM
Information provided by you is not sufficient to decide the approach that takes the least run-time and capable to grow over time. First, what is data type of KEY Variable(var1, var3, var5 and var8). What is the Maximum and Minimum of these in your 4 data sets(This has relevance to use ARRAYS)? My advice is to provide sample data sets(telling the data types) and show the output data set that you derive out of them. ... View more

Re: How to output observations whose data are missing for certain vari...

by Rhodochrosite | Level 12 in SAS Procedures
‎01-27-2015 11:59 AM
‎01-27-2015 11:59 AM
Let us take a small example based on SASHELP.CLASS. Create a data set, CLASS from SASHELP.CLASS and at the same time, get the number of numeric and character variables. Also make some the variables to be missing. data class;    set sashelp.class end = eof;    array n _numeric_;    array c _character_;    do _n_ = 1 to dim(n);       if ranuni(123) < 0.7 then n[_n_] = .;    end;    do _n_ = 1 to dim(c);       if ranuni(123) < 0.4 then c[_n_] = ' ';    end;    if eof then do;       call symputx('num', dim(n));       call symputx('chr', dim(c));    end; run; %put &num  ; %put &chr= ; Now the data set, CLASS will resemble what you start with. Now we will check each variable for MISSING, and if so, make it to take a label of 1(one) and if not missing then the label will be 0(zero). data want;    set class end = eof;    array n _numeric_;    array c _character_;    array miss_num _numeric_;    array miss_chr _character_;    do i = 1 to dim(n);       miss_num = ifN(missing(n), 1, 0);    end;    do i = 1 to dim(c);       miss_chr = ifN(missing(c), 1, 0);    end;    output; drop i;     run; The output data set, WANT, is what you are looking for ( as I understand your question). I realized that two arrays will do and not 4 arrays in the last program. The revised program follows: data want;    set class end = eof;    array n _numeric_;    array c _character_;    do i = 1 to dim(n);       n = ifN(missing(n), 1, 0);    end;    do i = 1 to dim(c);       c = ifC(missing(c), '1', '0');    end;    output; drop i;      run; Message was edited by: MUTHIA KACHIRAYAN ... View more

Re: NOBS= variable and DIM(array) as array subscript?

by Rhodochrosite | Level 12 in SAS Programming
‎01-27-2015 08:54 AM
3 Likes
‎01-27-2015 08:54 AM
3 Likes
Let us take a small example. Have the Data Set, CLASS, from SASHELP.CLASS. data class;    set sashelp.class; run; We wish to count the Frequency of AGE. We assume that the Maximum Age of the CLASS is 16. (This assumption is not needed as we could find the Max and Min Age in the previous step). We use an array K with a dimension of 16 as no AGE can exceed it in the data set. We contrive a trivial way to use the number observations of the data set so that we read one observation at a time using I as a counter  until the counter(I) does not exceed NUM which is the total number of observations of the Data Set. When I = NUM, the do-loop ends and the next do-loop, do i = 1 to dim(k); is carried out. The Subscript of the array K takes values from 1 to dim(k) which happens to be 1 to 16. In this do-loop, if the array-cell contains a value > 0, then it is output. The Output data set is shown below. Hope this helps you to understand what you seek. data want;    array k[16] _temporary_;    do i = 1 by 1 while ( i <= num);       set class nobs = num;       k[Age] + 1;    end;    do i = 1 to dim(k);       if k > 0 then do;          Age = i;          Freq = k;          output;       end;    end; keep Age Freq; run; Obs    Age    Freq 1      11      2 2      12      5 3      13      3 4      14      4 5      15      4 6      16      1 ... View more

Re: Select Most Important variables before a Linear Regression, Please...

by Rhodochrosite | Level 12 in SAS Data Science
‎01-26-2015 11:05 AM
‎01-26-2015 11:05 AM
[1] Examine the strength of correlation coefficient of variable i with the dependent variable. Say, choose r > 0.5 or some reasonable value. [2] Suppose you have chosen X1, X2, X3, ... X10. Check the linear relationship between each of them. Keep in your model only those that have lesser correlations( to avoid collinearity). [3] Explore this way until you can have manageable independent variables for your final model. ... View more

Re: Combining and filling in gaps in datasets

by Rhodochrosite | Level 12 in SAS Programming
‎01-26-2015 06:05 AM
‎01-26-2015 06:05 AM
When ARRAY can be used, HASH Object too can be used. Here is Hash version. BY GROUP processing with CLEAR() would be far easier as done with the ARRAY. No need to have BY GROUP processing. No need to change the date to number. Both Data Sets are read sequentially and the hash table is built. data one; input code $3. @9 date $5. count; datalines; 1X1     07/01         1 1X1     07/03         1 1X1     07/06         0 2Y3     .             0 3J8     07/02         1 3J8     07/03         0 ; run; data two; input code $3. @9 date $5. count; datalines; 1X1     07/01         0 1X1     07/02         1 1X1     07/04         0 2Y3     07/01         1 3J8     .             0 ; run; data want;    if _n_ = 1 then do;       if 0 then set one;       declare hash h(ordered:'a');       h.definekey('code','date');       h.definedata('code', 'date', 'count_14', 'count_13');       h.definedone();       declare hash h1(ordered:'y');       h1.definekey('code');       h1.definedone();       declare hiter hi('h1');    end;    do until(last);       set one(in = a) two(in = b) end = last;            if h.find() ^= 0 then do; count_14 = 0;  count_13 = 0; end;       if a then count_14 = count;       if b then count_13 = count;       if not missing(date) then h.replace();             if h1.find() ^= 0 then h1.add();    end;    do while(hi.next() = 0);       do Date = '07/01','07/02','07/03','07/04','07/05','07/06','07/07';          rc = h.find();          if rc ^= 0 then do;             count_14 = 0;             count_13 = 0;          end;          output;       end;    end; drop rc count;      run; ... View more

Re: Combining and filling in gaps in datasets

by Rhodochrosite | Level 12 in SAS Programming
‎01-25-2015 11:53 PM
‎01-25-2015 11:53 PM
No need for ETS. Simple data step is sufficient to get an elegant solution. Change Date into  a Number by using the MDY() function. Use one array for 2014(say, k_14[ ])  and anther array for average of 2012 & 13 (say, k_13[ ]). In double DoW-loops, read observations from the two data sets by BY GROUP (by CODE). Then merge the two arrays into one and output the results. data one(drop=date dd mm); input code $3. @9 date $5. count; mm = substr(date,1,2); dd = substr(date, 4); ndate = mdy(mm,dd, 2014); datalines; 1X1     07/01         1 1X1     07/03         1 1X1     07/06         0 2Y3     .             0 3J8     07/02         1 3J8     07/03         0 ; run; data two(drop=date dd mm); input code $3. @9 date $5. count; mm = substr(date,1,2); dd = substr(date, 4); ndate = mdy(mm,dd, 2014); datalines; 1X1     07/01         0 1X1     07/02         1 1X1     07/04         0 2Y3     07/01         1 3J8     .             0 ; run; data want(rename = (i = Date)); array k_14[19905:19911] _temporary_; array k_13[19905:19911] _temporary_;    do until(last.code);       set one;       by code;       if not missing(ndate) then k_14[ndate] = count;    end;    do until(last.code);       set two;       by code;       if not missing(ndate) then k_13[ndate] = count;    end;    do i = lbound(k_14) to hbound(k_14);       count_14 = ifN(k_14 > 0, k_14, 0);       count_13 = ifN(k_13 > 0, k_13, 0);       output;    end;    call missing(of k_14 , of k_13 ); keep code i count_14 count_13; format i date5.; run; ... View more

Re: replacing contents of a variable

by Rhodochrosite | Level 12 in Statistical Procedures
‎01-24-2015 09:03 AM
‎01-24-2015 09:03 AM
Not sure of your question. Do you want a solution like this? proc format; value num2ch 1 = 'Female'              2 = 'Male'              ; run; data have; input Sex; datalines; 1 1 2 2 1 2 ; run; data want;    set have;    format Sex num2ch. run; proc print data = want; run; proc contents data = want; run; But the type of SEX is still Number in WANT data set. ... View more

Re: Splitting dataset on total no of observations

by Rhodochrosite | Level 12 in SAS Enterprise Guide
‎01-22-2015 10:32 AM
‎01-22-2015 10:32 AM
Use FIRSTOBS = and OBS = Like: FIRSTOBS = 1 OBS = 1000000 Firstobs = 1000001 OBS = 2000000 Fiirstobs = 2000001 OBS = 3000000 ... View more

Re: Using hash to aggregate strings

by Rhodochrosite | Level 12 in SAS Programming
‎01-22-2015 08:23 AM
‎01-22-2015 08:23 AM
Actually you may not need the Hash. But with Hash, you need not SORT the data set before. See the code below. data _null_;    if _n_ = 1 then do;    length Text $100;       if 0 then set A;       declare hash h(ordered:'a');       h.definekey('Nr');       h.definedata('Nr','Text');       h.definedone();    end;    do until(last);       set A end = last;       if h.find() ^= 0 then Text = '';       call catx('|', Text, N);       h.replace();    end;    if last then h.output(dataset:'Want');    stop; run; ... View more

Re: Difference between dates

by Rhodochrosite | Level 12 in Statistical Procedures
‎01-21-2015 06:01 AM
‎01-21-2015 06:01 AM
Your CODE is Fine. Better to see your example. ... View more

Re: Find circular routes using Hash Objects

by Rhodochrosite | Level 12 in SAS Programming
‎01-19-2015 06:14 AM
‎01-19-2015 06:14 AM
Hi DeepH, For ID = 2, the recursive lookup works the way you want but not for ID = 1. In the former, you take From_City as KEY Part and get To_City as Data Part. Again take To_City(i) as KEY Part and get J as Data Part. And finally take J as Key Part and get I as the Data Part. In the latter, you will get a->b->c->e->f ... View more

Re: Find circular routes using Hash Objects

by Rhodochrosite | Level 12 in SAS Programming
‎01-19-2015 05:43 AM
‎01-19-2015 05:43 AM
DeepH, I have heard about recursive lookups (chained relationship) but not circular routes. For id = 2, can you explain how you get i->j->i. ... View more

Re: Count occurence of each word in a table

by Rhodochrosite | Level 12 in SAS Procedures
‎01-09-2015 12:21 AM
3 Likes
‎01-09-2015 12:21 AM
3 Likes
Yet another  simplified version of HASH Solution: data have; length word $8; input; nwords = countw(_infile_, ' '); do i = 1 to nwords;    word = scan(_infile_, i, ' ');    output; end; datalines; My cat is blue I like dogs The cat is blue ; run; data _null_;    if _n_ = 1 then do;       length word $8;       declare hash h();       h.definekey('word');       h.definedata('word', 'count');       h.definedone();    end;    do until(last);       set have end = last;       if h.find() ^= 0 then count = 0;       count + 1;       h.replace();    end;    if last then h.output(dataset:'W_Counts'); run; proc print data = w_counts; run; ... View more

Re: Replace Set-Merge loop with hash table lookup

by Rhodochrosite | Level 12 in SAS Procedures
‎01-07-2015 06:15 AM
3 Likes
‎01-07-2015 06:15 AM
3 Likes
Let ED2 contain SASHELP.CLASS. Let your Diagnosis be NAME. Another Data set, D1, has SEVEN NAMES instead of  14 diagnosis. The purpose is LOOKUP D1 for every NAME of ED2 and make the GROUP as you wanted. data ed2;    set sashelp.class; run; data d1; input Name $8.;         datalines;   Barbara James  Jane   Joyce  Louise Robert Ronald ; run; data want;    if _n_ = 1 then do;       if 0 then set d1;       declare hash h(dataset:'d1', hashexp:20);       h.definekey('Name');       h.definedone();    end;    set ed2;    Group = 'Unavoidable';    if h.find() = 0 then Group = 'Avoidable'; run; ... View more

Re: how to reverse a string with out using reverse function

by Rhodochrosite | Level 12 in SAS Programming
‎01-07-2015 04:50 AM
‎01-07-2015 04:50 AM
Reversing the String IN PLACE. data _null_;    t = 'Hello World SAS Communities';    length C $1;    len = length(t);    j = len;    do i = 1 to len while(i < j);       C = substr(t, i, 1);       substr(t, i, 1) = substr(t, j, 1);       substr(t, j, 1) = C;       j +- 1;    end;    put t = ; run; ... View more
Contact Me
Online Status
Offline
Date Last Visited
2 weeks ago
Group Hubs for KachiM
View All
Likes given to
Likes from
Quick links