Do sum of all corresponding values in one column when other column has a certain value

[Moaiz Siddiqui]

I want to sum the column 'stat1' by the column 'Group' if it has a certain value, such as Group == 'AAA'

Group	sub-group	code	Name	stat1	stat2	stat3
AAA	AA	wsed	A	2	5	20
		gfb	B	54	43	22
	BB	wewe	C	33	43	35
		qsq	D	4	20	43
BBB	CC	nhn	K	54	43	22
		asa	L	4	8	8
	DD	dsd	A	67	8	2
		ss	B	56	7	45

In non stata-syntax, I m looking for some command on the following lines:

Do sum of all corresponding values in column stat1 when value of Group == "AAA"

I would appreciate if some Stata user can help

[Nick Cox]

Consider the following silly example, which you can run. Here rep78 has the same role as Group in your example as I understand it.

sysuse auto, clear 
egen wanted = total(price), by(rep78) 
tabdisp rep78, c(wanted)

Also check out

help tabstat

However, note that Stata will not understand empty strings as meaning use the non-empty value above.

[Moaiz Siddiqui]

Thank you Mr. Nick Cox. This was very helpful.
I have to ask just one more small thing:

The code:
tabdisp rep78, c(wanted)


tabulates the variable 'rep78' in alphabetical order and the corresponding numerical values of the variable 'wanted' in the next column.

In case, I want to display 'rep78' in terms of numerical values of 'wanted' in descending order, how will the code change?
Will appreciate your help in this matter.

Regards,

Moaiz

[Nick Cox]

Here is one to do it using groups (Stata Journal):

. sysuse auto, clear 
(1978 Automobile Data)

. egen wanted = total(price), by(rep78) 

. groups wanted rep78, show(none) order(high) colorder(2) 

  +----------------+
  | rep78   wanted |
  |----------------|
  |     3   192877 |
  |     4   109287 |
  |     5    65043 |
  |     2    47741 |
  |     1     9129 |
  +----------------+