i have a time series data file that has about 3.5million observations. to give an example, the original data looks like this:
Column1: Column2: Column3:
x c1 t1
y c2 t2
. c3 t3
. c4 t4
in Column1, i have values X, Y, respectively in the first two rows. the rest of the rows in Column1 is empty. i want to use X and Y (values in the first two rows of Column1) to generate value for all the other rows in Column1, using the formula Column2 * (Column3 - lag2(column1)) + lag2(column1)
for example, to get the value for Column1, row 3, we want to take the lag2 value of the row, which is x, and do C3 *(t3-x)+x. for Column1, row 4, it should be C4 *(t4-y)+y
we want to finished data to be like this
Column1:
x
y
C3 *(t3-x)+x
C4 *(t4-y)+y
C3 *(t3-(C3 *(t3-x)+x))+(C3 *(t3-x)+x)
C4 *(t4-(C4 *(t4-y)+y))+(C4 *(t4-y)+y)
.
.
.
i have tried doing something like:
if row>2 thencolumn1 = lag2(Column1)
but it doesn't work.
Can you figure something out? Thanks!作者: shiyiming 时间: 2009-5-25 23:48 标题: Re: 利用一列row1里数据求下面row的数 --- 也想求一名sas 高手， 以后可以随时回到出现的一些问题。合理付费！ 我也遇到过这样的问题，跪求达人指教。作者: shiyiming 时间: 2009-5-26 10:05 标题: Re: 利用一列row1里数据求下面row的数 --- 也想求一名sas 高手， 以后可以随时回到出现的一些问题。合理付费！ 你试着去添加一个column，然后用这个column去记录行数。比如讲，这个column的名字叫obs。你就可以用：
if obs>2 then......作者: Qiong 时间: 2009-5-26 11:54 标题: Re: 利用一列row1里数据求下面row的数 --- 也想求一名sas 高手， 以后可以随时回到出现的一些问题。合理付费！ You may try the below code.
[code:ks1nnmin]data temp;
input y c t;
cards;
1 2 3
2 3 4
. 5 6
. 7 8
. 9 10
. 100 10
;
run;
data v;
retain la1 la2;
set temp;
x=lag(y);
if _n_=2 then do;
la2=x;
la1=y;
end;
if _n_>=3 then do;
y=c*(t-la2)+la2;
la2=la1;
la1=y;
end;
keep y c t;
run; [/code:ks1nnmin]