Counting observations for a household id split into multiple observations

[Saini Smriti]

Hello
I have a household data with a01 as the household id. Plotid represents the number of farming plot that household owns. h2_01 denotes the irrigation methods (labels below). I need to find out as to how many households are using rainfed as irrigation source (h2_01 = 1). One household can have multiple plots and can rely on rains for all those plots. I want to count the multiple plot observations as just "one" so as to get the number of farmers relying on rain.
I tried using contract and bysort command but that gives me the count for all plots of a household. Eg, if a01 = 4 has two different plotids (2 & 4) and it depends on rains (h2_01 =1 ) for both, I just want to count h2_01 for a01 = 4 only once. Please help.

* Example generated by -dataex-. To install: ssc install dataex
clear
input int a01 float plotid byte h2_01
 1   2 1
 1   3 3
 1   3 3
 3   2 2
 3   3 2
 3   5 1
 4   2 1
 4   4 1
 4   5 2
 4   7 2
 5   2 3
 5   2 3
 5   3 1
 6   2 3
 7   2 1
 9   2 3
10   3 6
10   7 3
end
label values h2_01 h2_01
label def h2_01 1 "rainfed", modify
label def h2_01 2 "river", modify
label def h2_01 3 "canal", modify
label def h2_01 4 "pond", modify
label def h2_01 6 "canal irrigation", modify
label def h2_01 7 "groundwater", modify

[Carlo Lazzaro]

Saini:
you may want to try something along the following lines:

. egen group=group( a01 h2_01) if h2_01==1


. egen tag=tag( group)

. replace tag= group if tag==1


. tab tag if tag!=0

 tag(group) |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |          1       20.00       20.00
          2 |          1       20.00       40.00
          3 |          1       20.00       60.00
          4 |          1       20.00       80.00
          5 |          1       20.00      100.00
------------+-----------------------------------
      Total |          5      100.00

.

[Saini Smriti]

Hello Carlo

That worked! I require a little more help.
I am trying to estimate the number of farmers who rely on rainfed over the total number of farmers in the dataset. As we can say, a01 == 5 does not rely on rainfed. There are multiple observations to each a01 (farmer). When I try to count the number of total farmers (bys a01 : egen farmer = count(a01)), it gives me the total by multiple obs (say, 3 for a01==5). I understand that I am using a wrong command.
Is there a way to use tag command to figure out the total number of farmers? Or any other command that helps?

Thanks
Smriti

[Saini Smriti]

Hey Carlo!

I figured it out via the tag command. Just removed the rainfed condition from there and used the same command to calculate the number of farmers.

Thanks
Smriti