Hi all,   I'm working with some tricky raw data and need to convert all numeric columns in an imported .csv file to character type.   I've followed this tutorial and got close to where I need to be, but I think something in the final data step is reverting the type of the columns back to numeric and I get the below note in the log: NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column). 35:3 Can anyone see anything in my code that might explain this or is able to explain the log note's meaning a bit further for me so I can find the issue myself.   My code looks like this: /* Import April 2019 form */ PROC IMPORT OUT= raw_new_neo_april19 DATAFILE= ***confidental*** DBMS=CSV REPLACE; GETNAMES=YES; DATAROW=2; GUESSINGROWS=2000; RUN; /* Create dataset with all column names and type */ proc contents data = raw_new_neo_april19 noprint out = vars1 (keep = name type); run; /* Use SQL to create macros with column names */ proc sql; * create macro list of all numeric variables names; select name into :numerics separated by ' ' from vars1 where type = 1; * create macro list of the numeric variables' names with the suffix "C"; select trim(name) || 'C' into :characters separated by ' ' from vars1 where type = 1; * create macro list of the conversion from the original name to the new name with the suffix "C"; select cats(name, ' = ' , name, 'C') into :conversions separated by ' ' from vars1 where type = 1; quit; /* Check column names all look okay */ %PUT &numerics; %PUT &characters; %PUT &conversions; /* use the macro list "&conversions" to add the suffix "C" to the numeric variables' names */ proc datasets library = work nolist; modify raw_new_neo_april19; rename &conversions; quit; /* Use arrays and put to convert all numeric columns to character columns */ data raw_new_neo_april19_2; set raw_new_neo_april19; array nums[*] &numerics; array chars[*] &characters; do i = 1 to dim(nums); nums[i] = put(chars[i],10.); end; drop i &characters; run; ... View more

Hello,   I'm using SGPlot to plot a bar chart from a dataset that gets updated once a week with a new observation for a total number of events that week.   The x-axis is getting crowded so I'd like to display the text label for every other axis tick, but keep all the bars. I've tried using the values= option but that removes the bars as well as the tick labels. I've also tried using the FITPOLICY=thin option, which is closer to what I want as it removes tick labels but keeps the data. But that doesn't keep enough of the x axis ticks and removes the most recent tick label which is most important.    Can anyone help me out?    proc sgplot data=set5 noborder noautolegend; yaxis label = 'Number' valueattrs=(size=10pt) LABELATTRS=(size=10pt) values=(0 to 100 by 5); xaxis label = 'Week' valueattrs=(size=7pt) LABELATTRS=(size=6pt) values=(1 to &max_week by 2); vbar week / datalabel datalabelfitpolicy=none datalabelattrs=(size=8) fillattrs=(color='CX0066CC'); format week donweek.; title1 h=12pt j=left "Figure 2"; keylegend/ title=" "; run; ... View more

I have two single column data sets with unique variables. I'd like to merge them to create a 2 column data set with an observation for every pair of variables from the the two original data sets. See example below.   I have the below two data sets with 1 column each: cat1 a b c d e   cat2 1 2 3 4   I want the below data set:   cat1 cat2 a 1 a 2 a 3 a 4 b 1 b 2 b 3 b 4 c 1 c 2 etc... etc... ... View more

Hi all,   I have a dataset of the form:   ID    var1    var2    var3 1       1         3          11 1       2         4          12 1       2         5        NUL 2      1          5          13 2      2          5         NUL 3      1          5         NUL 3      1          4         NUL   I'd like to remove duplicate observations of ID based on multiple criteria. For example: var3 should be non-null, var2 should be 5 and var1 should be 2. But if those criteria can't be met I want to keep the observation that best fits those criteria. e.g in the table given, I'd want to keep the 2nd observation with ID = 1 because var1 = 2 and var is non-null. I'd want to keep the first observation of ID=2 because var2 = 5 and var3 is non-null and I'd want to keep the first observation of ID=3 because var2 = 5.   Is this possible to solve with a SAS program, or is the original dataset too messy?   Thanks in advance for your help! ... View more