The DATASOURCE Procedure
- Overview
-
Getting Started
Structure of a SAS Data Set Containing Time Series Data Reading Data Files Subsetting Input Data Files Controlling the Frequency of Data – The INTERVAL= Option Selecting Time Series Variables – The KEEP and DROP Statements Controlling the Time Range of Data – The RANGE Statement Reading in Data Files Containing Cross Sections Obtaining Descriptive Information on Cross Sections Subsetting a Data File Containing Cross Sections Renaming Time Series Variables Changing the Lengths of Numeric Variables -
Syntax
-
Details
-
Examples
BEA National Income and Product Accounts BLS Consumer Price Index Surveys BLS State and Area Employment, Hours, and Earnings Surveys DRI/McGraw-Hill Format CITIBASE Files DRI Data Delivery Service Database PC Format CITIBASE Database Quarterly COMPUSTAT Data Files Annual COMPUSTAT Data Files, V9.2 New Filetype CSAUC3 CRSP Daily NYSE/AMEX Combined Stocks
- References
| LENGTH Statement |
The LENGTH statement, like the LENGTH= option in the ATTRIBUTE statement, specifies the number of bytes used to store values of variables in output data sets. The default lengths for numeric variables depend on the file type. Usually default lengths are set to 5 bytes.
The default lengths of character variables are defined as the minimum number of characters that can hold the longest possible value.
For some file types, the LENGTH statement also controls the amount of memory used to store values of numeric variables while processing the input data file. Thus, specifying LENGTH values smaller than the default will reduce both the disk space taken up by the output data sets and the amount of memory used by the PROC DATASOURCE step, at the cost of precision of output data values.
Any later length specification for a variable, using either the LENGTH statement or the LENGTH= option in the ATTRIBUTE statement, always overrides the previous one.