Fill backward a number of rows in SAS

Question

I need to create a variable that fills one cell 10 observations backward and one forward in SAS. A condition must be met. It is hard to explain, so the data below might help.

So far I tried to use proc expand to generate 10 leads and get them into one variable but that looks ridiculous and does not really work.

proc expand data=optionsreturns out=optionsreturns2;by secid cp_flag;

convert best_bid = ann_best_bid1 /transformout = (lead 1);
convert best_bid = ann_best_bid2 /transformout = (lead 2);
convert best_bid = ann_best_bid3 /transformout = (lead 3);
convert best_bid = ann_best_bid4 /transformout = (lead 4);
convert best_bid = ann_best_bid5 /transformout = (lead 5);
convert best_bid = ann_best_bid6 /transformout = (lead 6);
convert best_bid = ann_best_bid7 /transformout = (lead 7);
convert best_bid = ann_best_bid8 /transformout = (lead 8);
convert best_bid = ann_best_bid9 /transformout = (lead 9);
convert best_bid = ann_best_bid10 /transformout = (lead 10);
convert best_offer = ann_best_offer1 /transformout = (lead 1);
convert best_offer = ann_best_offer2 /transformout = (lead 2);
convert best_offer = ann_best_offer3 /transformout = (lead 3);
convert best_offer = ann_best_offer4 /transformout = (lead 4);
convert best_offer = ann_best_offer5 /transformout = (lead 5);
convert best_offer = ann_best_offer6 /transformout = (lead 6);
convert best_offer = ann_best_offer7 /transformout = (lead 7);
convert best_offer = ann_best_offer8 /transformout = (lead 8);
convert best_offer = ann_best_offer9 /transformout = (lead 9);
convert best_offer = ann_best_offer10 /transformout = (lead 10);

run;

the Option Contract Price   P=Put   of the Option   ID  Bid Across All  Ask Across All  Vol Average if Negative)    Price   WANT    COUNT
10006273    19980303    C   19980418    5168    9.875   10.125  0   34.6875 .       
10006273    19980304    C   19980418    5168    9   9.25    0   33.8125 .       
10006273    19980305    C   19980418    5168    9   9.25    0   33.75   .   0.25    -10
10006273    19980313    C   19980418    5168    12.375  12.75   0   37.3125 .   0.25    -9
10006273    19980331    C   19980418    5168    19.625  20.125  0   44.625  .   0.25    -8
10012764    19960105    C   19960120    5168    0.375   0.4375  71  30.5    .   0.25    -7
10012764    19960108    C   19960120    5168    0.1875  0.375   0   30.1875 .   0.25    -6
10012764    19960109    C   19960120    5168    0.0625  0.1875  0   29.375  .   0.25    -5
10012764    19960110    C   19960120    5168    0.125   0.25    0   28.75   .   0.25    -4
10012764    19960111    C   19960120    5168    0   0.125   15  28.875  .   0.25    -3
10012764    19960112    C   19960120    5168    0   0.125   0   28.375  .   0.25    -2
10012764    19960115    C   19960120    5168    0   0.125   0   28.5    .   0.25    -1
10012764    19980824    C   19960120    5168    0.25    0.4375  28  29.25   24/08/1998  0.25    0
10022220    19960205    C   19960420    5168    18.75   19.125  26  33.625  .

Forget about the one forward. Just when date=ann_date then fill from 0 to -10 backward — MaBo88, Aug 22 '19 at 10:08
This is the same as [Generate multiple lags through loops in SAS?](https://stackoverflow.com/questions/57456363/generate-multiple-lags-through-loops-in-sas/57456656#57456656) — Richard, Aug 22 '19 at 10:40
Possible duplicate of [Generate multiple lags through loops in SAS?](https://stackoverflow.com/questions/57456363/generate-multiple-lags-through-loops-in-sas) — Reeza, Aug 22 '19 at 15:03

Fill backward a number of rows in SAS

0 Answers0