Stata Technical Bulletin
Options
start (date-literal) is required and specifies the starting date for the first observation in the dataset. dateJiteral takes forms
such as 1964, 1999ml, 1960ql, 12janl985, and so on, depending on whether the data is monthly, quarterly, daily, and
so on; see [U] 27 Commands for dealing with dates for more on how to specify dates.
sequenceVaanamee) specifies that verneme contains an integer variable that specifies the sequence of the observations. This
allows gaps to be specified for the time variable. If the values of varname are not sequential, the resulting time variable
will have gaps.
Note that the start date specified in start () is taken to be the date of the value of the sequence () variable in the first
observation. If that value is missing, then the date from start() is associated with a value of 1 for the sequence() variable.
Description
tsmktim creates a Stata time variable, newtimeear, with an appropriate format for yearly, twice yearly, quarterly, monthly,
weekly or daily data, and executes tsset to use that variable as the time specifier. Note that the data must be ordered by time
before issuing tsmktim.
If the sequence() option is not specified, the data are assumed to have no gaps and are sequential in the periodicity of
dateJiteral, that is, one quarter after another, or one month after another, with no gaps.
Examples
Assume the following data:
X
44
21
15
77
Then
. tsmktim mytime, start(1977q2)
will produce
x mytime
44 1977q2
21 1977q3
15 1977q4
77 197Sql
while
. tsmktim mytime, start(29decl948)
will produce
x mytime
44 29decl94S
21 30decl94S
iε 31decl948
77 01janl949
and
. tsmktim mytime, start(29decl948) sequence (myseq)