Stata/Excel time formatting issue - excel

Question

I have a variable minutes that I am having trouble cleaning/standardizing.
It is imported from Excel in a Date Time format, but I just want the minutes. For example, if a player plays 30 minutes and 34 seconds, it's displayed as 30:34 or 30:34:00. However, it's stored in Excel as 6:34 AM (30:34 is interpreted as military time), or 12:34 AM, depending on whether it is 30:34 or 30:34:00. Thus it ends up getting imported into Stata as 6:34 or 12:34, when the value I want is what's displayed (30:34 or 30:34:00). Is there a way for you to format a number in Excel to just be the value that is displayed?
Once it's imported into Stata it's impossible to standardize, because you cannot differentiate a player that plays 30:34 (when it displays 30:34:00) from a player that plays 6:34 (they will both show 6:34).

Answer 1

Please make a greater effort when posting questions. People in capacity to help might ignore the question because it is difficult to understand, because you provide no code (and thus show no effort), because the problem is not reproducible, and more.
Suppose an MS Excel sheet like the following
Then the following should get you started:
clear
set more off
import excel timetest.xls, cellrange(C2:C4) firstrow
gen hour = hh(time)
gen sec = mm(time)
gen realmin = hour + 24
order realmin, before(sec)
list
resulting in
. list
+-------------------------------------------+
| time hour realmin sec |
|-------------------------------------------|
1. | 01jan1900 06:34:00 6 30 34 |
2. | 01jan1900 00:00:00 0 24 0 |
+-------------------------------------------+
See help datetime. This is a must if working with dates and/or time.
Note that adding 24 won't work for any data set.
A general solution takes the form of
clear
set more off
import excel timetest.xls, cellrange(C2:C5) firstrow
gen hour = hh(time)
gen sec = mm(time)
gen t = dofc(time)
format t %td
gen dayselap = t - td(01jan1900) + 1
gen realmin = hour + (24 * dayselap)
drop hour t dayselap
order time realmin
list
For example, this data in MS Excel:
time
30:34:00
24:00:00
58:04:00
65:00:00
will produce
. list
+------------------------------------+
| time realmin sec |
|------------------------------------|
1. | 01jan1900 06:34:00 30 34 |
2. | 01jan1900 00:00:00 24 0 |
3. | 02jan1900 10:04:00 58 4 |
4. | 02jan1900 17:00:00 65 0 |
+------------------------------------+
(There might be an issue with leap years that you are encouraged to research yourself.)
An alternative solution may involve converting the values to text (within MS Excel) and then managing the text within Stata.

Answer 2

I am not at all familiar with Stata.
The problem is that the actual data in Excel is being entered incorrectly. Excel stores dates and times as days + fractions of a day, and then displays it according to the formatting.
So:
30:34 is really 0 hours 30 minutes 34 seconds and is stored as 2.1226851851851854E-2 which is the computation: 30/(24*60)+34/(24*60*60)
However 30:34:00 is really 30 hours, 34 minutes 0 seconds and is stored as 1.273611111111111 which is the computation 30/24+34/(24*60)
Depending on whether it is entered properly as 0:30:34, or improperly as 30:34 determines the value that is stored.
In Excel, you could possibly pre-process the values if you have some test to tell if the data was entered correctly or not.
For example:
=IF(time_unreasonable,A1/60,A1)
would convert the values if they had been incorrectly entered as hours:minutes rather than minutes:seconds.
Other solutions may also be feasible, perhaps even in Stata.