Help on Stata Data cleaning and analysis

[Sera Sarkar]

I have a data set with company names and their daily returns value. I want to calculate the information share month-wise for which I first need to clean the data. Following are the issues I am facing and need help with:
1) I want to drop the company from the dataset if it has a missing return value for even a single day. Example: for the following table, since a, d has missing ret, a and d will both be dropped from the data set completely(even the ones with ret values)

Date	Company names	ret
1st Jan 2020	a	1
1st Jan 2020	b	2
1st Jan 2020	c	3
1st Jan 2020	d
2nd Jan 2020	a
2nd Jan 2020	b	1
2nd Jan 2020	c	2
2nd Jan 2020	d	3
3rd Jan 2020	a	1
3rd Jan 2020	b	2
3rd Jan 2020	c	3
3rd Jan 2020	d	1
4th Jan 2020	a
4th Jan 2020	b	1
4th Jan 2020	c	2
5th Jan 2020	a	3
5th Jan 2020	b	1
5th Jan 2020	c	2
6th Jan 2020	d	3
6th Jan 2020	a	3

2) I want to calculated the information share if company c(suppose) has 19 returns in jan'2020 and 19 returns in the subsequent month (feb'2020) i.e. equal number of returns . I was planning to use a dummy variable to differentiate . like dummy variable =0 for equal count of ret for jan and feb and 1 otherwise. How to go about this? Any alternate suggestions for this.

P.S: I am new to Stata and appreciate your help. Thanks a lot

[Nick Cox]

1) is

bysort company (ret) : drop if missing(ret[_N])

I don’t follow what you want otherwise.

[Sera Sarkar]

Thank You for the response.

for 2) I want to create a dummy variable for Jan returns. The dummy variable will be 0 if the number of returns for a company x is equal to the number of returns of the same company in the subsequent month and 1 otherwise. Hope I am clear with the query.
PS: I have already created a t variable for the time series such that Jan 2020 is grouped as 1, Feb 2020 a as 2 etc.
Thank you again.

[Sera Sarkar]

Hi Nick Cox , With reference to the code you provide, what will it look like if I want to run this code for each value of t (my timeseries variable)?

[Nick Cox]

#4 That would just be equivalent to

drop if missing(ret)

if I understand correctly.

(I still don't understand #3.)