Reading Raw Data

Reading Column-Binary Data

Definition

column-binary data storage: is an older form of data storage that is no longer widely used and is not needed by most SAS users. Column-binary data storage compresses data so that more than 80 items of data can be stored on a single "virtual" punched card. The advantage is that this method enables you to store more data in the same amount of space. Because card-image data sets remain in existence, SAS provides informats for reading column-binary data. See Description of Column-Binary Data Storage for a more detailed explanation of column-binary data storage.

How to Read Column-Binary Data

To read column-binary data with SAS, you need to know:

how to select the appropriate SAS column-binary informat
how to set the RECFM= and LRECL= options in the INFILE statement
how to use pointer controls.

The following table lists and describes SAS column-binary informats.

SAS Informats for Reading Column-Binary Data
Informat Name		Description
$CBw.		reads standard character data from column-binary files
CBw.		reads standard numeric data from column-binary files
PUNCH.d		reads whether a row is punched
ROWw.d		reads a column-binary field down a card column

To read column-binary data, you must set two options in the INFILE statement:

Set RECFM= to F for fixed.
Set the LRECL= to 160, because each card column of column-binary data is expanded to two bytes before the fields are read.

For example, to read column-binary data from a file, use an INFILE statement in the following form before the INPUT statement that reads the data:

infile file-specification or path-name
  recfm=f lrecl=160;

Note: The expansion of each column of column-binary data into two bytes does not affect the position of the column pointer. You use the absolute column pointer control @, as usual, because the informats automatically compute the true location on the doubled record. If a value is in column 23, use the pointer control @23 to move the pointer there. [cautionend]

Description of Column-Binary Data Storage

The arrangement and numbering of rows in a column on physical punched cards originated with the Hollerith system of encoding characters and numbers. It was based on using a pair of values to represent either a character or a numeric digit. In the Hollerith system, each column on a card had a maximum of two punches, one punch in the zone portion, and one in the digit portion. These punches corresponded to a pair of values, and each pair of values corresponded to a specific alphabetic character or sign and numeric digit.

In the zone portion of the punched card, which is the first three rows, the zone component of the pair can have the values 12, 11, 0 (or 10), or not punched. In the digit portion of the card, which is the fourth through the twelfth rows, the digit component of the pair can have the values 1 through 9, or not punched.

The following figure shows the multi-punch combinations corresponding to letters of the alphabet.

Columns and Rows in a Punched Card

[Columns and Rows in a Punched Card]

SAS stores each column of column-binary data (a "virtual" punched card) in two bytes. Since each column has only 12 positions and since 2 bytes contain 16 positions, the 4 extra positions within the bytes are located at the beginning of each byte. The following figure shows the correspondence between the rows of "virtual" punched card data and the positions within 2 bytes that SAS uses to store them. SAS stores a punched position as a binary 1 bit and an unpunched position as a binary 0 bit.

Column-Binary Representation on a "Virtual" Punched Card

[Column-Binary Representation on a Punched Card]

Top of Page