display format

[tim bradshaw]

I have a numeric variable that contains some large numbers (up to 10 digits long).
The variable format is %12.0g. I can see the full-length number in the data view window but when I tabulate this variable, in the results window it shows the shorthand version (e.g. 5.04e+8), instead of the full-length number.
Is there a way to view the full-length number in the results window when running a tab command?
Many thanks

[Nick Cox]

No example data or precise commands here.

Instead of tabulate, consider table, tabdisp, list, etc. Which might be good will depend on the details you don't give.

[tim bradshaw]

Hi Nick, thanks for your reply.

For example:

year	id
2000	123456789
2000	987654321

format id %12.0g
tab id

produces...

id | Freq. Percent Cum.
------------+-----------------------------------
1.23e+08 | 1 50.00 50.00
9.88e+08 | 1 50.00 100.00
------------+-----------------------------------
Total | 2 100.00


...whereas I would like it to display:

id | Freq. Percent Cum.
------------+-----------------------------------
123456789 | 1 50.00 50.00
987654321 | 1 50.00 100.00
------------+-----------------------------------
Total | 2 100.00


Any advice on this would be greatly appreciated!
Thanks

[tim bradshaw]

Ah, 'table' did the trick - thank you!

[tim bradshaw]

Actually, now there is the same problem with 'table' but in the frequency count column.

[Nick Cox]

Thanks for the example.

tab id

works fine for me so long as you first change the display format of id to something like

format id %10.0f

[tim bradshaw]

Got it, thanks

[tim bradshaw]

Hi again Nick, do you find that negative numbers are still shortened in the results window when using the following command?

 format id %10.0f

tab id

[Nick Cox]

As before, please give data examples explicitly. Here groups (Stata Journal) seems a possibility.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float year double id
2000 -123456789
2000 -987654321
end

format id %12.0f

tab id

         id |      Freq.     Percent        Cum.
------------+-----------------------------------
  -9.88e+08 |          1       50.00       50.00
  -1.23e+08 |          1       50.00      100.00
------------+-----------------------------------
      Total |          2      100.00

groups id

  +---------------------------------------+
  |         id   Freq.   Percent      %<= |
  |---------------------------------------|
  | -987654321       1     50.00    50.00 |
  | -123456789       1     50.00   100.00 |
  +---------------------------------------+

[Sergiy Radyakin]

Tim,

1. tabulating IDs is rarely a productive task. If you have 10-digit IDs then you probably have a lot of them, resulting in some mega-tables, not digestible by humans.

2. consider collapsing your dataset by ID, then listing the resulting dataset.

3. another possibility is to generate myid=string(id) with some format, then tabulate myid. Here you have full control over how you create a string (readable) representation of ID.

4. avoid tabulate command as much as possible and use table command instead.

clear
input float year double id
2000 -123456789
2000 -987654321
end

expand 100000

format id %20.0gc
table id, format(%20.0gc)

format id %20.0g
table id, format(%20.0g)

Output

. format id %20.0gc
. table id, format(%20.0gc)
-----------------------------------
          id |                Freq.
-------------+---------------------
-987,654,321 |              100,000
-123,456,789 |              100,000
-----------------------------------
. format id %20.0g
. table id, format(%20.0g)
---------------------------------
        id |                Freq.
-----------+---------------------
-987654321 |               100000
-123456789 |               100000
---------------------------------

Best, Sergiy

[dr mmr]

Same probelem is being faced when using table to do sum of a numeric varibale across a catrorical variable
See 2.73e+07
How to het the actual number diplayed
Tried all combinations in varibale propoerties (Long/double)

[dr mmr]

Region sum(ER4)

ARO 2385917
AWR 56968
EN 1039412
ESEAOR 2.73e+07
SARO 376872
WHR 774809

[Nick Cox]

Please study the Statalist FAQ

https://www.statalist.org/forums/help#realnames

https://www.statalist.org/forums/help#question

https://www.statalist.org/forums/help#stata

I can't say that you have a question that isn't already answered in this thread.

A display format like %8.0f will show numbers like 27312345 with all digits shown.

[Marcelo Dell]

I am having the same problem.
I can display the variable in the format I want but the output of the statistical tests using the variable (i.e. ttest) is in the imprecise format.

[Rich Goldstein]

re: #14 - it is not clear what part of the output you are complaining about, but see

help pformat