Reading variables from a external file (csv) that have different number of observations

[Wagner_Alvarenga]

Suppose the following csv file (named "raw"), with 3 variables (A, B, C):

A  B  C

2  5  6

4  7  5

5  0  2

6  2  4

.  9  3

.  8  5

.  5  9

.  .   7

I want to read this information and save on a data set named "raw_data_readed":

data raw_data_readed;

  infile ' c:\raw.csv ' dlm= ',' ;

  input var1 var2 var3;

run;

But when SAS reads, the result is:

obs   var1 var2 var3

1       2     5     6

2       4     7     5

3       5     0     2

4       6     2     4

Is there a way to have obs 1 to 8?

Although here I told that there are only 3 variables, in real there are more variables.

[ballardw]

Your example data is not in CSV format. If the data is actually csv and no informat has been associated with variables I would expect to get missing values for Var1 for obs 5-8 and missing for var2 on obs 8 but not completely missing records.

You may need to post your actual reading code if not a few lines of data.

[Wagner_Alvarenga]

Ballardw,

I sent a file.

I would appreciate if you may help me!

[Wagner_Alvarenga]

For the file that I sent, how should I do tho have the following variables:

Data Set TEST (with 4 variables var1, var2, var3, var4)

2	5	6	2
4	7	5	9
5	0	2	8
6	2	4	5
.	9	3	.
.	8	5	7
.	5	9	0
.	.	7	.

[ballardw · Accepted solution]

The DSD option is what you were looking for. Here is an example reading datalines instead of an external file so folks can see whats going on.

data raw_data_readed;

  infile datalines  dlm= ',' dsd ;

  input var1 var2 var3 var4;

datalines;

2,5,6,2

4,7,5,9

5,0,2,8

6,2,4,5

,9,3,

,8,5,7

,5,9,0

,,7,

run;

[Wagner_Alvarenga]

Thanks Ballardw!!