How do I refer to the current line/observation number in a loop in the data step?

0
votes

For example, I have data on various latencies that I iterate through and apply a linear function to, as shown: (the function is just an example here)

data latency;
input lat1 - lat20;
array cost[20];
array lat[20];
do x = 1 to 20;
cost[x] = lat[x] * 1.875;
end;
drop x;
datalines;
0.42 0.85 0.59 0.06 0.21 0.35 0.1 0.08 0.85 0.53 0.81 0.44 0.47 0.2 0.99 0.32 0.18 0.87 0.33 0.84
0.11 0.83 0.02 0.59 0.74 0.65 0.76 0.45 0.57 0.22 0.2 0.13 0.42 0.15 0.05 0.51 0.48 0.95 0.39 0.92
0.8 0.9 0.65 0.29 0.77 0.0 0.24 0.05 0.16 0.72 0.58 0.9 0.35 0.63 0.79 0.41 0.73 0.36 0.82 0.16
0.74 0.21 0.57 0.73 0.83 0.78 0.77 0.92 0.13 0.39 0.52 0.14 0.1 0.77 0.68 0.99 0.26 0.37 0.97 0.83
;
run;

How can I save a variable with the current observation number in each iteration of the loop, so that I can use it in calculations later?

I know that proc print will automatically print the observation number, but how do I access this and store it to a variable in the data step? Is there a way to do this as sas reads the datalines line by line?

I tried this, but then the obs variable is 2 for every observation. 

data latency;
input lat1 - lat20;
obs = 1;                 * ADDED LINE;
array cost[20];
array lat[20];
do x = 1 to 20;
cost[x] = lat[x] * 1.875;
end;
obs = obs + 1;           * ADDED LINE;
drop x;
datalines;
0.42 0.85 0.59 0.06 0.21 0.35 0.1 0.08 0.85 0.53 0.81 0.44 0.47 0.2 0.99 0.32 0.18 0.87 0.33 0.84
0.11 0.83 0.02 0.59 0.74 0.65 0.76 0.45 0.57 0.22 0.2 0.13 0.42 0.15 0.05 0.51 0.48 0.95 0.39 0.92
0.8 0.9 0.65 0.29 0.77 0.0 0.24 0.05 0.16 0.72 0.58 0.9 0.35 0.63 0.79 0.41 0.73 0.36 0.82 0.16
0.74 0.21 0.57 0.73 0.83 0.78 0.77 0.92 0.13 0.39 0.52 0.14 0.1 0.77 0.68 0.99 0.26 0.37 0.97 0.83
;
run;

proc print data=latency;
run;

This is a small example, but in reality I can't simply add a new variable that stores the line number to the start of each data line and read it in. That isn't practical for the actual data set.

2
sas

2 Answers

2
votes

You just need to add a retain statement so SAS doesn't reset obs to 0 at every new observation.

data latency;
retain obs 0;
obs = obs + 1;
...
run;
1
votes

Your first attempt was very close. Try again, but this time replace this line:

obs = 1;                 * ADDED LINE;

With this:

retain obs 0;            * ADDED LINE;

That way, your obs variable will be retained across your entire dataset instead of being reset to 1 each time.