How to manage missing values in my dataset?

[Lucas Bordure]

Good morning.
I have 10 columns from A to J [Year, Industry, Ticker, EVA, Size, Leverage, growth, Por, Volatility, Intangible assets)
In my column EVA, Size, Leverage, growth, Por, Volatility, Intangible assets, I have missing values (#VALUE!, #N/A N/A, #DIV/O!)
How to manage this in Stata?

[Ken Chui]

Stata uses a period (.) as the default missing. There are auto manually defined missings that users can choose, ranging from .a to .z, 26 of them. Unless the users need to identify the types of missing, most of the time just using the default one is sufficient.

You can use -browse- in Stata to open a spreadsheet view. The black numbers are numeric and they seem to be corrected imported, and if you scroll down you'll see some cells with just a period in it, those are the default missing.

You can also see some red numbers, those are string variables (akin to character) and currently cannot be used as numbers. Those "#N/A N/A" were interpreted as character, causing it to be imported as text. And you may recover the numerical data using the codes:

gen growth_num = real(GROWTH)
gen intangible_asset_num = real(INTANGIBLEASSETS)

or:

destring GROWTH, gen(growth_num2) force
destring INTANGIBLEASSETS, gen(intangible_asset_num2) force

To learn more about missing in Stata, check out: https://www.stata.com/manuals/u12.pd...1Missingvalues and https://www.stata.com/manuals/dmissingvalues.pdf

[Lucas Bordure]

Thank you for your reply.
I can see period (.) in many columns right now but I still observe #N/A N/A for GROWTH and INTANGIBLEASSETS.

[Ken Chui]

I can see period (.) in many columns right now but I still observe #N/A N/A for GROWTH and INTANGIBLEASSETS.

Of course you do, they are still their old selves. The new numerical versions of them are the two new varaibles "growth_num" and "intangible_asset_num". If you plan to do any statistical summery, use these two new ones.

As for GROWTH and INTAGIBLEASSETS, there is no point to replace those "NAs" into empty. That will not change their nature of being a string variable which is unsuitable for many statistical summary/analysis.

Just as an example, try run:

mean GROWTH
mean growth_num

If for any reason you plan to stick to the old names, then delete the old variable, and give that name to the new one. For example:

drop GROWTH
rename growth_num GROWTH

[Lucas Bordure]

Ok. Thank you a lot for your explanations Ken Chui. You helped me a lot!
Have a nice day.