Hello,   I am working on a map of the United States with CBSA lines set in SAS EG (I think v9?). This is really my first time trying to tweak with the code other than changing patterns/colors/titles etc. so I don't have much experience to work off of. Basically, I am trying to use the CBSA codes I have in combination with the CBSA locations that are drawn on the map to make a new dataset that has each CBSA code and ALL of the CBSAs that it is contiguous to (borders touch) in a horizontal dataset.   Example scenario. There are four CBSAs, 10000,10001,10002,10003. 10003 touches all the others, 10002 touches 10003 and 10001, 10000 only touches 10003. From this, I want a dataset that looks like this:   MAIN_CBSA      CONT_CBSA1   CONT_CBSA2   CONT_CBSA3 10000                 10003 10001                  10002                  10003 10002                  10001                  10003 10003                  10000                  10001                   10002     The on idea I had was to take the coordinates dataset I have and try to take observations that have the same x and y values and write wide the cbsa code for each but I don't know how to do this.   data that I have:   x          y             cbsa_name              cbsa_code 1          2               New York                 10001 1          2               Brooklyn                   10002   data I want: x          y            cbsa_name               cbsa_code                 cont_cbsa1 1          2              New York                   10001                        10002 1          2              Brooklyn                     10002                        10001     I just don't know how to generate variables for each observation based off of the x and y.   The order of the cont_cbsa* variable doesn't matter. I just need them all in a row, because eventually I will be pulling a count of patients (for medicare) across each of the CBSAs and want to know whether a sum of the main_cbsa and any of the others equals 500+. If this isn't doable that's totally fine I just thought I'd give it a shot.   Thanks ... View more

Hello,   I am trying to output a table from a proc freq based on a grouping variable, a binary indicator, and the year. The output automatically stores the three variable values and the count for each, but I would also like the column (or row) percent to also be outputted because eventually this will be made into a formatted table in excel where I need those.   The data that the proc freq is being done on looks like this.   id        group          indicator              year 11          1                    0                      2012 11           1                   1                      2013 16           1                   0                      2012 18           1                   1                      2012 22          5                    1                      2011   The output table currently looks like: group               indicator             year            count 1                            0                   2012              2 1                            1                   2012              1 1                            1                   2013              1 5                            1                   2011              1   And I want the column percent added from the proc freq printed output: group               indicator             year            count       _column_pct_ 1                            0                   2012              2                 66.67 1                            1                   2012              1                 33.33 1                            1                   2013              1                 100 5                            1                   2011              1                 100   The trouble I am having is typically when I do these proc freqs with only two tables it automatically outputs the column percentage, but when I try 3 variables it doesn't do it. I still want the same column percent that is being shown in the results window though.   Thanks.   ... View more

Hello,   I have a dataset with a bunch of addresses, unfortunately they are all combined into one line and I need to parse out everything that comes after the person's name into one line.   Data that I have:   id         Name               address_line                                              city_state         zipcode 1          John Doe          john doe 123 Central Avenue                 New York, NY    12345   Data I want: id         Name               address_line_1                                          city_state         zipcode 1          John Doe         123 Central Avenue                                 New York, NY    12345     I know I probably have to use index and substring or some combination of those, but I really have no idea where to start. ... View more

I am sure this is easy, but I have been trying to figure it out and can't. I have a list of IDs, where some of them are duplicates. What I want is to number all of the ids so that they have a number next to them identifying which observation is which based on some sort criteria.   Data:   ID       Numeric_Var 1                 5 2                28 2                42 3                  2 3                  9 3                 74     And I want a new variable so that my data looks like this:   ID       Numeric_Var         Count 1                 5                        1 2                28                       1 2                42                       2 3                  2                       1 3                  9                       2 3                 74                      3   Any help would be great.   Thanks ... View more

Hello,   I am trying to put together a bunch of partial proc freq tables and would like to do this automatically if possible. What I have is a bunch of proc freqs that are cross tabs of two different binary indicator variables. An example is:   Table of iah by prov iah prov Frequency Percent Row Pct Col Pct 0 1 Total 0 168177 34.08 34.15 99.67 324327 65.73 65.85 99.88 492504 99.81     1 553 0.11 59.08 0.33 383 0.08 40.92 0.12 936 0.19     Total 168730 34.19 324710 65.81 493440 100.00     From this (and about 20 more) I want to grab the bolded red numbers from each of the 20 tables to make a new dataset/table that I can print and have to output to excel. I just don't know how to reference these values.   Thanks ... View more

Hi,   I am running using proc logistic to get a propensity score to then match individuals from groups to create comparison groups. In the proc logistic output I want to ONLY see the Analysis of MLE table because I am eventually only going to export that to Excel.   These are the current specifications I have:   proc logistic data=test_a descending; class female; model test=p_total_ab age65_older female / SELECTION=stepwise RISKLIMITS LACKFIT RSQUARE PARMLABEL; output out=propensity_scores prob= prob; run;   Again, the question I have is how to isolate the Analysis of MLE table to only output that. I understand noprint doesn't print anything, but I want the MLE table.   ... View more

My data looks like this.   Obs   Id   Numeric       Character 1        1       1                  xxx 2        1       2                  xxx 3        2       3                  xyz 4        2        4                 xyz 5        3        5                 zzz 6        3         6                zzz   What I want to do is create a new table that is grouped by the ID variable, while summing the numeric variable and retaining the value of the character variable (which is the same for every ID).   I know how to do this for this exact example with proc sql, a select statement and summing Numeric, and using a group by statement, but the problem is there are hundreds of numeric variables in my actual data set and I want to do it for every one. Anyone have any ideas?   Thanks ... View more

So I am trying to do a proc freq for data that looks like this   id_num   race xxx            1 xxx            2 xxx            3 xxx            4   So when I do proc freq on race I want to the display to show actual races instead of 1,2,3,4 like white, black, asian, hispanic.   I have tried a few things (replacing values but that didn't work at all) but can't get anything close so all I really have is.   proc freq data=test; tables race; title; footnote; run;   which brings up Race Code RACE Frequency Percent Cumulative Frequency Cumulative Percent 0 3772 0.98 3772 0.98 1 185936 48.20 189708 49.18 2 139248 36.10 328956 85.28 3 9606 2.49 338562 87.77 4 13978 3.62 352540 91.39 5 27078 7.02 379618 98.41 6 6127 1.59 385745 100.00 Frequency Missing = 16     but I ... View more

I am trying to add unique titles to my outputs for proc means and proc univariate results, but the only way I know how to do this is through the insertion of a macro variable which can only be one word. I could use "_" but it needs to be cleaner. Anyone know of a way to create unique titles within a macro process?   What I have so far looks like this:   data have; set hadbefore; run;   %let varA= abc def ghi ; %let varB= j    k     l m  n    o p   q    r ; %macro test; %do i=1 %to 3; ods noproctitle; %let v = %scan(&varA, &i.); %let t = %scan(&varB, &i.); title "&v."; title2 "&t."; proc means data=test;     (another similar program with proc univariate as well) var &v; run; %end; %mend test; %test;   What it does for the  title 2 is prints one word for each instead of the whole line. I have tried adding in quotations but then it runs with an error because it picks up the first word with one quotation mark.   Any help would be great thanks. ... View more