String or not to String on Stata: Collection of usefull tools

Stata ENG

Random string utilities that saved me in the past.

Parsing cmdline after a estimation command.

. qui reg price rep78 length foreign , robust 

. local what_I_typed = e(cmdline)

. di "`what_I_typed'"
regress price rep78 length foreign , robust

gettoken: be carefull with the spaces (!!)

we can split the command (regress) from the rest.

. gettoken command rhs : what_I_typed ,parse(" ,") // first token and the rest

. di "`command'"
regress

. di "`rhs'"
 price rep78 length foreign , robust

Also we can split the explained variable (price) from all the rest.

. gettoken y x_plus_options : rhs ,parse(" ,") // first token and the rest

. di "`x_plus_options'"
 rep78 length foreign , robust

Finally we can parse the regressors from the options using parse(“,”) instead of parse(" ,"). Yes, they are different.

. gettoken x options : rhs ,parse(",") // split using "," (look carefully!)

. di "`x'"
price rep78 length foreign 

. di "`options'"
, robust

word() function

Get the first and last regressors