cond() function : how to generate a dummy variable, but not put == 1 to missing values?

[Michael Duarte Goncalves]

Hi,

I want to create a variable using -cond()- as follows:

gen tariff_2_more_15000_w = cond(tariff_2 == 1 & (power_p1 <=15000 | power_p2 <=15000), 1,0)
replace tariff_2_more_15000_w = 0 if tariff_2 == 1 & (power_p1 >15000 | power_p2 >15000)

* Example generated by -dataex-. For more info, type help dataex
clear
input long(power_p1 power_p2) float(tariff_2_more_15000_w tariff_2)
 3300     . 0 1
 3450  3450 1 1
 2300  2300 1 1
 3450  3450 1 1
 3450  3450 1 1
 2300  2300 1 1
 5400  5400 1 1
 3300  3300 1 1
 3450     . 0 1
 2300  2300 1 1
 1725     . 0 1
 3450  3450 1 1
 3450  3450 1 1
 3100  3100 1 1
 1000  1000 1 1
 3300  3300 1 1
 2000  2000 1 1
 2300  2300 1 1
 3450  3450 1 1
 3450  3450 1 1
 4600  4600 1 1
 1725  1725 1 1
 3450     . 0 1
    .     . 0 1
 2300  2300 1 1
 3100  3100 1 1
 3450  3450 1 1
 6200  6200 1 1
 2200  2200 1 1
 2300  2300 1 1
 2000  2000 1 1
 3500  3500 1 1
 2300  2300 1 1
 2300  2300 1 1
 2300  2300 1 1
 2300  2300 1 1
 2300  2300 1 1
 2300  2300 1 1
 3300  3300 1 1
 3450  3450 1 1
 3450  3450 1 1
 3450  3450 1 1
 3450  3450 1 1
 1150  1150 1 1
 5500  5500 1 1
 2300  2300 1 1
 3450  3450 1 1
 3450     . 0 1
 3450  3450 1 1
 2300  2300 1 1
 3100  3100 1 1
 3450  3450 1 1
 3450     . 0 1
 3450  3450 1 1
 3450  3450 1 1
 3100  3100 1 1
 3450  3450 1 1
 5500  5500 1 1
 3300  3300 1 1
 3300  3300 1 1
 2300  2300 1 1
 4000  4000 1 1
 2200     . 0 1
 2300  2300 1 1
 3500  3500 1 1
 3450  3450 1 1
 3450  3450 1 1
 3100  3100 1 1
 3400     . 0 1
 3450  3450 1 1
 2800  2800 1 1
 3450  3450 1 1
 3450  3450 1 1
 2300  2300 1 1
 2000  2000 1 1
 3100  3100 1 1
 2300  2300 1 1
13200 13200 1 1
 3450  3450 1 1
 2200     . 0 1
 3450  3450 1 1
 3300  3300 1 1
 2300  2300 1 1
 6200  6200 1 1
 2300  2300 1 1
 3300  3300 1 1
 3400  3400 1 1
 2300  2300 1 1
 2300  2300 1 1
 3450  3450 1 1
 2300  2300 1 1
 3450  3450 1 1
 3450     . 0 1
 3100  3100 1 1
 3450     . 0 1
 3450  3450 1 1
 2300  2300 1 1
 3000  3000 1 1
 3450  3450 1 1
 3450     . 0 1
end

However, the new variable contains also a dummy equals to one when both power_p1 and power_p2 are missing, which I don't want.
Basically, if one of variable power_p1 or power_p2 is missing, it should contain 1.

But, if both are missing, it should take the value of 0, instead of 1 right now.

Thank you for your help.
Best,

Michael

[Daniel Schaefer]

However, the new variable contains also a dummy equals to one when both power_p1 and power_p2 are missing, which I don't want.

I don't think so. Check your example again. If either power_p1 or power_p2 is missing, the expression evaluates to zero. I think this is what you want:

replace tariff_2_more_15000_w = (missing(power_p1) | missing(power_p2)) & !(missing(power_p1) & missing(power_p2)) ///
                                                       if tariff_2 == 1 & (missing(power_p1) | missing(power_p2))

[Daniel Schaefer]

Don't these two lines:

gen tariff_2_more_15000_w = cond(tariff_2 == 1 & (power_p1 <=15000 | power_p2 <=15000), 1,0)
replace tariff_2_more_15000_w = 0 if tariff_2 == 1 & (power_p1 >15000 | power_p2 >15000)

Turn out to be equivalent to this line?

gen tariff_2_more_15000_w = tariff_2 == 1 & (power_p1 <=15000 & power_p2 <=15000)

[Michael Duarte Goncalves]

Hi Daniel Schaefer,

It works well, thanks for the suggestion.
However, I would like to have for my variable tariff_2_more_15000_w a "." in lieu of 0 when both power_p1 and power_p2 are missing.

Here is a -dataex- example:

* Example generated by -dataex-. For more info, type help dataex
clear
input long(power_p1 power_p2) float(tariff_2 tariff_2_more_15000_w)
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
345000 345000 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
     .      . 1 0
end

Thank you in advance.
Best,
Michael

[Daniel Schaefer]

There are a few ways to do this, but the simplest in my mind is just:

replace tariff_2_more_15000_w = . if missing(power_p1) & missing(power_p2)

[Michael Duarte Goncalves]

Hi Daniel Schaefer,

Thank you for the suggestion. It is straightforward.

Best,

Michael