LOSTCARD Statement

Resynchronizes the input data when SAS encounters a missing or invalid record in data that has multiple records per observation.

Valid in:	DATA step
Categories:	Action
Categories:	CAS
Type:	Executable

Table of Contents

Syntax

LOSTCARD;

Without Arguments

The LOSTCARD statement prevents SAS from reading a record from the next group when the current group has a missing record.

Details

When to Use LOSTCARD

When SAS reads multiple records to create a single observation, it does not discover that a record is missing until it reaches the end of the data. If there is a missing record in your data, the values for subsequent observations in the SAS data set might be incorrect. Using LOSTCARD prevents SAS from reading a record from the next group when the current group has fewer records than SAS expected.

LOSTCARD is most useful when the input data have a fixed number of records per observation and when each record for an observation contains an identification variable that has the same value. LOSTCARD usually appears in conditional processing such as in the THEN clause of an IF-THEN statement, or in a statement in a SELECT group.

When LOSTCARD Executes

When LOSTCARD executes, SAS takes several steps:

Writes three items to the SAS log: a lost card message, a ruler, and all the records that it read in its attempt to build the current observation.
Discards the first record in the group of records being read, does not write an observation, and returns processing to the beginning of the DATA step.
Does not increment the automatic variable _N_ by 1. (Normally, SAS increments _N_ by 1 at the beginning of each DATA step iteration.)
Attempts to build an observation by beginning with the second record in the group, and reads the number of records that the INPUT statement specifies.
Repeats steps 1 through 4 when the IF condition for a lost card is still true. To make the log more readable, SAS prints the message and ruler only once for a given group of records. In addition, SAS prints each record only once, even if a record is used in successive attempts to build an observation.
Builds an observation and writes it to the SAS data set when the IF condition for a lost card is no longer true.

Example: Resynchronizing Input Data

This example uses the LOSTCARD statement in a conditional construct to identify missing data records and to resynchronize the input data.

data inspect;
   input id 1-3 age 8-9 #2 id2 1-3 loc 
         #3 id3 1-3 wt;
   if id ne id2 or id ne id3 then
    do;
      put 'DATA RECORD ERROR: ' id= id2= id3=;
      lostcard;
    end;
   datalines;
301    32
301    61432
301    127
302    61
302    83171
400    46
409    23145
400    197
411    53
411    99551
411    139
;

The DATA step reads three input records before writing an observation. If the identification number in record 1 (variable ID) does not match the identification number in the second record (ID2) or third record (ID3), a record is incorrectly entered or omitted. The IF-THEN DO statement specifies that if an identification number is invalid, SAS prints the message that is specified in the PUT statement message and executes the LOSTCARD statement.

In this example, the third record for the second observation (ID3=400) is missing. The second record for the third observation is incorrectly entered (ID=400 while ID2=409). Therefore, the data set contains two observations with ID values 301 and 411. There are no observations for ID=302 or ID=400. The PUT and LOSTCARD statements write these statements to the SAS log when the DATA step executes.

  
DATA RECORD ERROR: id=302 id2=302 id3=400
NOTE: LOST CARD.
RULE:----+----1----+----2----+----3----+----4----+----5----+----
14   302    61
15   302    83171
16   400    46
DATA RECORD ERROR: id=302 id2=400 id3=409
NOTE: LOST CARD.
17   409    23145
DATA RECORD ERROR: id=400 id2=409 id3=400
NOTE: LOST CARD.
18   400    197
DATA RECORD ERROR: id=409 id2=400 id3=411
NOTE: LOST CARD.
19   411    53
DATA RECORD ERROR: id=400 id2=411 id3=411
NOTE: LOST CARD.
20   411    99551

The numbers 14, 15, 16, 17, 18, 19, and 20 are line numbers in the SAS log.