Trying to replace a variable with a new generated variable

[Joseph Hopkin]

Hello,
I am trying to replace a variable so that when I create a graph for GDPpc it shows GDPpc below 20k and above 30k. the starting gdp variable is gdppc
I used: 'generate code_gdppc2 = "" ' to make a string variable, then:
replace code_gdppc2 = code_gdppc if gdppc < 20000

However, when this is run it says 0 real changes made and I cannot for the life of me figure out why.

or to get my desired result I run:

'replace code_gdppc2 = code_gdppc if gdppc < 20000 & gdppc > 30000'

and this says 'code_gdppc ambiguous abbreviation'
what is the way around this?

Apologies if this is classed as an elementary question, I am new here and have no other place to turn for an answer.
Thank you.

[Ali Atia]

A data example would be useful here - it's not really clear to me what the data looks like or what's causing the problem (particularly the code_gdppc variable). You can provide a data example by using the -dataex- command in Stata and pasting the output into the forum between [code] delimiters.

[Joseph Hopkin]

I am not sure either.....
I was told to create the new variable as 'code_gdppc' when learning

also I am sorry but I am not sure what you mean by using the -dataex- command. stata tells me this is unrecognized.
the mean for gdppc of 160 obs is 32365.64. I hope this can help, but I assume this is not what you mean by an example.

sorry for the evident noobiness

[Ali Atia]

If you're using a version of Stata before 14.2 you need to install dataex from SSC before you can use it, which you can do by typing:

ssc install dataex
dataex

[Joseph Hopkin]

ok thanks, i got this for the variable gdppc:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float gdppc
 31010.56
36186.297
37873.516
13442.896
19053.975
35269.254
 38247.39
 23425.33
37555.773
 72325.46
 22882.49
37472.348
 49313.43
 39066.07
39050.383
 23017.41
16351.788
 41825.84
 32085.58
 82480.42
 44289.63
 13192.96
 23559.57
26223.055
34603.145
 19588.87
 29734.87
21193.676
 49121.97
33087.715
24531.113
 50069.02
34787.105
 65324.16
 56206.96
 45591.65
 39669.91
 43958.76
 35340.15
  21151.3
  51812.8
 40328.89
34996.746
 42293.39
 27335.42
35118.273
25898.336
 56311.71
17579.547
20998.406
 45320.19
12292.755
 27110.67
 47665.27
29699.033
37056.695
 29992.85
 26163.98
 25710.05
 29587.87
 30648.85
 27895.58
 36210.06
27064.465
24460.375
 29922.96
 39644.54
 57050.86
11016.222
31167.344
34729.613
 49661.71
 34514.97
 39063.94
 45442.02
35837.906
 39468.19
17534.592
 33848.24
 43279.79
  26053.2
26672.535
27994.623
33271.848
 37300.19
 26032.73
 24125.22
 29883.59
14958.278
 20790.46
 22634.11
 22519.94
11500.647
24179.855
34059.563
23543.295
 59323.56
24968.955
 19269.07
 25329.38
end

[Ali Atia]

Could you also show the values of the code_gdppc variable which you're trying to replace code_gdppc2 with?

[Joseph Hopkin]

Maybe this is where I am going wrong, as code_gdppc is a string and just gives rows of: " " when I run the dataex on it.
" "

[Nick Cox]

The difficulty is over (abbreviation of) variable names, So, the results of

describe code_gdppc*

should be given.

https://www.statalist.org/forums/help#stata explains about data examples in general and dataex in particular.

while https://www.statalist.org/forums/help#version explains that you should tell us about an out-of-date version.

[Ali Atia]

Okay, so now your problem makes more sense. When you use the command:

replace code_gdppc2 = code_gdppc if gdppc < 20000 & gdppc > 30000

You're replacing one variable which is full of empty strings "" (code_gdppc2) with another variable which is full of empty strings (code_gdppc). So there are no changes made because the variables are already identical. Also, it's impossible for gdppc to be both below 20000 and above 30000 at the same time, so the condition of gdppc < 20000 & gdppc > 30000 will never be fulfilled.

I'm not sure what you're trying to achieve with the string variables, but if what you want is a variable which only contains gdppc when gdppc is lower than 20000 or higher than 30000, you can just do:

gen wanted = gdppc if gdppc < 20000 | gdppc > 30000

This variable will be missing when gdppc is between 20000 and 30000.

[Joseph Hopkin]

Okay this helps thank you. I have now created the desired graph but the 'wanted' mlabel results in the actual gdp number being put on the graph and not the name of the country which corresponds to that gdp at that point in time.
Do you know how I can mlabel the country name within these ranges and not their gdp?

Thanks

[Joseph Hopkin]

The difficulty is over (abbreviation of) variable names, So, the results of

describe code_gdppc*

should be given.

https://www.statalist.org/forums/help#stata explains about data examples in general and dataex in particular.

while https://www.statalist.org/forums/help#version explains that you should tell us about an out-of-date version.

Thanks Nick Cox, I will be sure to refer to these resources in future statalist posts.

[Nick Cox]

Unless you have solved your problem showing us the results of

 
 describe code_gdppc*

remains a concrete suggestion for here and now.