PROC IMPORT in SAS is used to read data to SAS. Reading data from an external file is the most frequent task of a SAS programmer.
One thing to remember is that PROC IMPORT procedure can import data only that data type is supported by SAS. SAS can import only numeric and character data types.
PROC IMPORT Syntax:
PROC IMPORT DATAFILE=filename OUT=sas-dataset<data-set-options> DBMS= data-source-identifier REPLACE; SHEET= sheet-name GETNAMES=Yes/No DATAROW=n; RANGE=range-values RUN;
DATAFILE=filename: It is used to specify the complete path and filename or fileref for the input file. A fileref is a SAS name that is associated with the physical location of the output file.
Note: A pathname for a file can have a maximum length of 201 characters.
OUT= SAS data-set: This tells SAS to create a SAS dataset. By default, SAS dataset is created in your work library. You can also specify a two-level SAS dataset name.
DBMS=data-source-identifier: This is an optional argument and is used to specify the type of data to import. To import a DBMS table, specify DBMS = using a supported database identifier.
For example, to import a tab-delimited file you must specify TAB as the identifier. If you have a delimited file that does not end in .csv specify the DLM as the identifier and the
delimiter= option. The default delimiter is blank.
The below table shows some of the values that are frequently used for the DBMS identifier.
DBMS Identifiers Supported in Base SAS
|Identifier||Output Data Source||Extension|
|CSV||Delimited file (comma-separated values)||.csv|
|DLM||Delimited file (other than blank)||.dat or .txt|
|TAB||Delimited file (tab-delimited values)||.txt|
|XLS||Excel 97-2003 workbooks||.xls|
|XLSX||Microsoft Excel 2007 and later||.xlsx|
|ACCESS||Microsoft Access 2000 and later||.mdb|
You can find the Other DBMS here in the SAS Help Center.
REPLACE: This is also an optional argument. If you specify the REPLACE option, it overwrites an existing SAS data set.
If you do not specify REPLACE, the IMPORT procedure and the dataset already exists, the PROC IMPORT procedure does not overwrite the existing dataset and a NOTE is written in the log with the below message:
NOTE: Import cancelled. Output dataset dataset-name already exists. Specify REPLACE option to overwrite it.
SHEET: When you have to import data from a specified sheet in excel, you can use this option to specify the sheet name. By default, SAS will import sheet1 from excel.
GETNAMES: SAS imports the first row of an excel sheet as the Variable name for the SAS variable. Similarly, if you specify No option, it will tell SAS, not to use the first row of data as variable names and instead, SAS creates variable names as VAR1, VAR2 and so on.
DATAROW= Using this option you can specify the starting row from where SAS would import the data. If you omit this option, SAS will import data starting from the 1st row of excel.
An important point to note here is:
When GETNAMES=YES, DATAROW must be greater than or equal to 2.
When GETNAMES=NO, DATAROW must be greater than or equal to 1
RANGE= For specifying the range of rows and columns of an excel sheet, use the range option with the sheet range as arguments.
Importing a Delimited File
The below code snippets is used to import a delimited .txt file and the records are separated by ‘|’. Note the use of
FirstName|LastName|Gender|Country|Age Dulce|Abril|Female|United States|32 Mara|Hashimoto|Female|Great Britain|25 Philip|Gent|Male|France|36 Kathleen|Hanner|Female|United States|25 Nereida|Magwood|Female|United States|58
proc import datafile="/home/9to5sas/examples/data.txt" dbms=dlm out=mydata replace; delimiter='|'; getnames=yes; run; proc print data=mydata; run;
Note: To import a tab-delimited file specify the delimiter option as below.
Importing a Tab-Delimited File into SAS
The below code snippet is an example of importing a Tab-delimited file in SAS. Note the delimiter=’09’x;
PROC IMPORT DATAFILE= "/home/9to5sas/examples/data.txt" OUT= mydata replace DBMS=dlm; delimiter='09'x; GETNAMES=YES; RUN;
Importing a Space-Delimited File
To import a space-delimited file , specify delimiter = ’20’x
PROC IMPORT DATAFILE= "/home/9to5sas/examples/data.txt" OUT= mydata replace DBMS=dlm; delimiter='20'x; GETNAMES=YES; RUN;
Importing a Comma-Delimited File with TXT extension
To import a comma-separated file that has a txt extension, specify delimiter = ‘,’
PROC IMPORT DATAFILE= "/home/9to5sas/examples/data.txt" OUT= mydata replace DBMS=dlm; delimiter=','; GETNAMES=YES; RUN;
Importing a Specific Delimited File Using a Fileref
When you use a fileref to specify a delimited file to import, the logical record length (LRECL) defaults to 256, unless you specify the
LRECL= option in the FILENAME statement. The maximum LRECL that the PROC IMPORT procedure supports is 32767.
FirstName LastName Gender Country Age Dulce Abril Female 'United States' 32 Mara Hashimoto Female 'Great Britain' 25 Philip Gent Male France 36 Kathleen Hanner Female 'United States' 25 Nereida Magwood Female 'United States' 58
filename mydata2 '/home/9to5sas/examples/data2.txt' lrecl=100; proc import datafile=mydata2 dbms=dlm out=mydata2 replace; getnames=yes; run; proc print data=mydata2; run;
Importing a Comma-Delimited File with a CSV Extension
By default PROC IMPORT procedure recognizes .csv as an extension for a comma-separated file so if you are importing a .csv file DBMS option is not required. However, it is required if you are importing a.txt file that has comma-delimited data.
proc import datafile="/home/9to5sas/examples/mydata3.csv" out=mydata3; run; proc print data=mydata3; run;
Importing a file containing multiple delimiters
If two or more delimiters are present in the input file, quote each of the delimiters following the delimeter= option.
PROC IMPORT DATAFILE= "/home/9to5sas/examples/mydata.txt" OUT= outdata DBMS=dlm REPLACE; delimiter=','09'x '; GETNAMES=YES; RUN;
Importing records from a specified row
You can tell SAS to start reading from a specified row using the
Here is an example.
proc import datafile=mydata2 out=mydata4 dbms=dlm replace; datarow=5; run; proc print data=mydata4; run;
Data will be read from row 5 due to the DATAROW= option.
Importing variable names other than the first row
Suppose you have variable names starting from the 2nd row in an excel file. In this case, you can use DATAROW= and STARTROW=
variable variable variable variable variable First Name Last Name Gender Country Age Dulce Abril Female United States 32 Mara Hashimoto Female Great Britain 25 Philip Gent Male France 36 Kathleen Hanner Female United States 25 Nereida Magwood Female United States 58
proc import datafile="/home/subhroster20070/examples/mydata5.xls" out=mydata5 dbms=xls replace; namerow=2; startrow=3; run; proc print;
NAMEROW=2 tells SAS to read variable names from the second row and STARTROW=3 is used to read values from starting from the third row. You need to use both the options otherwise the variable names will also be read in the 2nd observation.
Note: NAMEROW only works with XLS and not with XLSX format. For XLSX formats you can use the RANGE= option discussed later in this post.
Using the RANGE= option
To import a specific range of values, you can use the RANGE= option as below.
proc import datafile="/home/9to5sas/examples/data8.xls" out=mydata5 replace dbms=xls; range="data8$a15:e28" ; getnames=no; run; proc print;
Using the GUESSINGROWS= option
For delimited files, SAAS scans the first 20 rows are scanned to determine the variable type and length attributes. You can increase the number of rows that are scanned by using the GUESSINGROWS= statement.
Here is a scenario.
Suppose you have first 30 rows as numeric then the remaining are character values. SAS would make all the character values blank. To avoid this you can use specify the GUESSINGROWS=100 option.
See the column last name and Age in the below input file.
Row Number 29 has a character value in the Age field and Row Number 35 has a length more than the first 20 rows.
While importing this file, SAS writes an error message in the log. This is because SAS has determined the datatype of Age as Numeric by scanning the first 20 rows. And there is a character variable in the Numeric field. Hence the error.
Also, see the output below where there is a missing value at observation 28 and at observation 34, the value of the last name variable has been truncated.
To avoid these error and truncation of values, use the guessing rows option with a value of more than the rows in your data or use guessingrows=Max.
Note that if your input data contains thousands of records it would make the import process slow. Here, I have used
proc import datafile="/home/9to5sas/examples/data8.csv" out=mydata5 replace; guessingrows=100; run; proc print;