Statistical software SAS
Basic SAS programming
a online material to learn basic data handling and manipulation in SAS
(under construction)
Frequently asked questions:
How to use SAS import in SAS 9.4. (Error message: Connection failed).
Logistic regression:
In PROC LOGISTIC why aren't the coefficients consistent with the odds ratios?
Typical errors in SAS programming
Missing semicolons
The most typical error in SAS programming is to miss out on a semicolon in the end of the line. An error is given, but for SAS it is difficult to identify the problem as a missing semicolon. Below the error is indicated as the run; line. The error is in the line above. Run could then not be identified.
308 data indata1a;
309 set indata1;
310 if concent>25
311 run;
---
22
ERROR 22-322: Syntax error, expecting one of the following: !, !!, &, *, **, +, -, /, <, <=, <>, =, >, ><, >=, AND, EQ, GE, GT, IN, LE,
LT, MAX, MIN, NE, NG, NL, NOTIN, OR, ^=, |, ||, ~=.
Creating new data sets
Another typical error when making subsets or creating new variables in a data set is that either one of the names is misspelled or a condition is not fullfilled. In the log-window you see that the newly created data set indata1a has 0 observations. Observe that this is only given as a note not as error or warning.
A condition is not fullfilled:
NOTE: There were 48 observations read from the data set WORK.INDATA1.
NOTE: The data set WORK.INDATA1A has 0 observations and 3 variables.
The variable is misspelled:
NOTE: Variable concentr is uninitialized.
NOTE: There were 48 observations read from the data set WORK.INDATA1.
NOTE: The data set WORK.INDATA1A has 0 observations and 4 variables.