How to get results from first date with multiple labels?

[ddchina]

Dear All,

I have one data set. There are three scanners and each has two time points measurements with four different label results. I would like to get only the first time point results from these three scanners. But I can only get the first label results if I use the first option. Here is my code and data set I have and results I would like to have.

proc sort data=have; by Manufacturers study_date label ;run;

data want;

   set have;

   by Manufacturers study_date  label;

   if first.study_date;

run;

what I have:

Manufacturers	Study_Date	Label	mhpost	mhant	mhmid	mhup	mhlow
N85148	8-Jan-13	L1	67.25	65.88	61.50	82.05	82.81
N85148	8-Jan-13	L2	68.48	72.60	62.50	84.14	88.10
N85148	8-Jan-13	L3	69.83	80.63	67.58	84.59	87.05
N85148	8-Jan-13	L4	67.60	79.55	70.60	89.19	93.35
N85148	19-May-13	L1	67.34	65.46	61.58	82.39	83.10
N85148	19-May-13	L2	68.04	72.94	63.80	85.36	87.90
N85148	19-May-13	L3	69.00	76.90	66.82	86.65	88.14
N85148	19-May-13	L4	67.50	79.44	72.12	88.96	92.19
N85257	1-Apr-13	L1	67.25	66.25	61.80	81.19	82.42
N85257	1-Apr-13	L2	70.45	74.35	65.30	84.00	85.05
N85257	1-Apr-13	L3	70.50	78.80	66.20	86.85	89.84
N85257	1-Apr-13	L4	66.25	78.90	73.30	88.09	92.12
N85257	6-Jun-13	L1	66.80	66.32	61.44	82.37	82.78
N85257	6-Jun-13	L2	68.16	74.08	63.20	84.30	86.04
N85257	6-Jun-13	L3	68.78	77.94	65.84	86.81	88.15
N85257	6-Jun-13	L4	66.86	78.60	72.02	88.62	92.43
N85286	21-Feb-13	L1	67.78	67.00	63.35	81.55	81.90
N85286	21-Feb-13	L2	68.83	73.15	64.58	84.70	85.91
N85286	21-Feb-13	L3	69.55	78.08	66.73	85.19	86.96
N85286	21-Feb-13	L4	67.70	80.35	70.80	88.23	92.32
N85286	20-Apr-13	L1	66.52	66.10	61.58	81.19	82.04
N85286	20-Apr-13	L2	67.26	72.10	62.46	84.77	86.14
N85286	20-Apr-13	L3	69.46	76.96	65.78	84.74	87.23
N85286	20-Apr-13	L4	67.82	78.72	71.80	87.42	90.34

what I want:

Manufacturers	Study_Date	Label	mhpost	mhant	mhmid	mhup	mhlow
N85148	8-Jan-13	L1	67.25	65.88	61.50	82.05	82.81
N85148	8-Jan-13	L2	68.48	72.60	62.50	84.14	88.10
N85148	8-Jan-13	L3	69.83	80.63	67.58	84.59	87.05
N85148	8-Jan-13	L4	67.60	79.55	70.60	89.19	93.35
N85257	1-Apr-13	L1	67.25	66.25	61.80	81.19	82.42
N85257	1-Apr-13	L2	70.45	74.35	65.30	84.00	85.05
N85257	1-Apr-13	L3	70.50	78.80	66.20	86.85	89.84
N85257	1-Apr-13	L4	66.25	78.90	73.30	88.09	92.12
N85286	21-Feb-13	L1	67.78	67.00	63.35	81.55	81.90
N85286	21-Feb-13	L2	68.83	73.15	64.58	84.70	85.91
N85286	21-Feb-13	L3	69.55	78.08	66.73	85.19	86.96
N85286	21-Feb-13	L4	67.70	80.35	70.80	88.23	92.32

Thank you so much for your help in advance,

Bo

[PGStats · Accepted solution]

You have to sort by label before study_date :

proc sort data=have; by Manufacturers label study_date;run;

data want;

   set have;

   by Manufacturers label;

   if first.label;

run;

PG

[ddchina]

Thank you so much for your prompt help PGStats! It works! Bo

[ballardw]

Or assuming at sorted by manufacturers

data want (drop= temp:);

set have;

by Manufacturers ;

length tempman $ 40;

retain tempman tempdate;

if first.Manufacturers then do;

tempman=Manufacturers;

tempdate=study_date;

end;

if tempman=Manufacturers and

   tempdate=study_date;

run;

[ddchina]

Thank you Ballardw for your suggestion! Bo