adding a new obs to existing dataset

[thomask23]

I need to add additional observations Balamt=0 where a the previous month (Monyear) value is null for a cateogy,

Would the best approach be to create  new table(see below) and then do a Union query?

data test;

   input monyear $ val $ Category $;

   datalines;

mar-15 0 1 mkt1

mar-15 0 2 mkt2

mar-15 0 1 mkt7

;

or could i create a if then data step to add the Balamt -0 if the previous months data is null?

[Steelers_In_DC]

Here  you go, hope this helps:

data have;

    infile datalines dsd;

    informat monyear2 date9.;

    format monyear2 date9.;

   input monyear $ monyear2 val $ Category $;

   datalines;

,31MAR2015,0 1,mkt1

mar-15,,0 2,mkt2

mar-15,,0 1,mkt7

;

run;

data want;

set have;

/*if using characters*/

if monyear = '' then balamt = '0';

/*if using numbers*/

if monyear2 = . then balamt2 = 1;

run;

[ballardw]

It will help if you can show some example data from the base data set and what the desired output for the example data would be.

[thomask23]

below is an example of the current dataset and the desired data

current table

BALAMT MONYEAR Segment

100           Mar-15      agncy7

200           Mar-15      agncy2

300           Mar-15      agncy3

400           Mar-15      agncy1

100           Mar-15      sports4

needed table  

BALAMT MONYEAR Segment

0                Feb-14      agncy1

400           Mar-15         agncy1

0                Feb-14        agncy2

200           Mar-15           agncy2

0                Feb-14           agncy3

300           Mar-15           agncy3

0                Feb-14           agncy7

100           Mar-15           agncy7

0                Feb-14           sports4

100           Mar-15           sports4

[Steelers_In_DC]

I see what you meant about the union now, I misunderstood what you were asking for.  Yes I would build a data set with the date range first and then join.  I've used proc expand for this as well.

[Astounding]

If this is really an exact representation of the input and output, you could try:

data want;

   set have;

   output;

   balamt=0;

   monyear='Feb-14';

   output;

run;

Sorting might be necessary afterwards.  Good luck.