RE: st: ado file help

[tashi lama <ltashi32@hotmail.com>]

Works like a magic. Thanx guys. There is a slight problem however. If I were to extend this and include varlist in the syntax, 

......
syntax [varlist], STARTdate(str) ENDdate(str) BROKERid(numeric) GRAPHtype(str)
obdc load, exec("select date(read_date), count(*) from readership where source_id==`brokerid' and date(read_date) between '`startdate'' and '`enddate'' group by date(read_date)")
.......
end 

and when I call 
dd var1, startdate("2011-01-01") enddate("2011-01-20") broker(6429) 
stata isn't happy and returns "no variables defined". It makes kind of sense syntax command doesn't see any data since it is defined before generating dataset. However, I need to define syntax, options  to generate dataset in first place. Any idea how to solve this? 

Thanx tons 

-----------------------------------
> Date: Tue, 14 Aug 2012 15:40:05 +0100
> Subject: Re: st: ado file help
> From: njcoxstata@gmail.com
> To: statalist@hsphsun2.harvard.edu
>
> There are broadly two ways to learn a language, (a) to try it out and
> (b) to read about it. Both are needed, but (a) is not working well for
> you because you are not giving (b) a fair try.
>
> 1. Your -syntax- is illegal if only because -numeric- is _not_ allowed
> as a keyword by itself. Either you want to specify an integer argument
> or a real one. See my email at
> http://www.stata.com/statalist/archive/2012-08/msg00396.html for a
> recent reminder.
>
> 2. Your program says that -brokerid()- is an option, but you typed
> -brokerod()-. There is another typo in -graphttype()-.
>
> 3. Also, the purpose of capitalisation in a -syntax- statement is to
> signal acceptable abbreviations of option names. It doesn't indicate
> how those option names should be typed when invoking a program
> yourself.
>
> Nick
>
> On Tue, Aug 14, 2012 at 3:26 PM, tashi lama <ltashi32@hotmail.com> wrote:
> > Ok. The confusion I had was my arguments were either varlist/namelist/anything and hence I was reading and I quote from Stata "If you type nothing, the command doesn't allow a varlist/namelist/anything". In any case, I had
> > .....
> > syntax,STARTdate(str) ENDdate(str) BROKERid(numeric) GRAPHtype(str)
> > ...
> > obdc load, exec("select date(read_date), count(*) from readership where source_id==`BROKER_id' and date(read_date) between '`STARTdate'' and '`ENDdate'' group by date(read_date)")
> > if "`GRAPHtype'"=="pie" {
> > ......
> > ..
> > end
> >
> > and when I run
> > dd, STARTdate("2011-01-01") ENDdate("2011-01-20") BROKERod(6429) GRAPHTtype("pie")
> > stata isn't happy and throws back invalid syntax.
> >
> > ----------------------------------------
> >> Date: Tue, 14 Aug 2012 01:42:08 +0100
> >> Subject: Re: st: ado file help
> >> From: njcoxstata@gmail.com
> >> To: statalist@hsphsun2.harvard.edu
> >>
> >> Daniel's advice is excellent. More directly, answers are
> >>
> >> 1. No.
> >> 2. No.
> >>
> >> You don't need varlist, namelist, or anything as the four things you
> >> want to specify on the command line are all definable as arguments to
> >> options. From your own question that would be sufficient.
> >>
> >> Nick
> >>
> >> On Mon, Aug 13, 2012 at 11:01 PM, tashi lama <ltashi32@hotmail.com> wrote:
> >> > Sorry Nick, this doesn't help much.
> >> > 1. Do we not need to define varlist/namelist/anything after syntax
> >> > 2. How do you put it as macro in the query. Is it going to be
> >> > syntax anything, STARTdate(str) ENDdate(str) BROKERid(numeric) /* anything because the arguments are not varlists
> >> > odbc load, exec("select date(read_date), count(*) from readership where source_id=`BROKER_id' and date(read_date) between '`STARTdate'' and '`ENDdate'' group by date(read_date)")
> >> > or
> >> > odbc load, exec("select date(read_date), count(*) from readership where source_id=`anything list' and date(read_date) between '`anything list'' and '`anything list'' group by date(read_date)")
> >> >
> >> > ----------------------------------------
> >> >> Date: Mon, 13 Aug 2012 22:25:46 +0100
> >> >> Subject: Re: st: ado file help
> >> >> From: njcoxstata@gmail.com
> >> >> To: statalist@hsphsun2.harvard.edu
> >> >>
> >> >> As you want to write a program, you will find it much easier to write
> >> >> it in Stata style
> >> >>
> >> >> program dd
> >> >> syntax , STARTdate(str) ENDdate(str) BROKERid(str) GRAPHtype(str)
> >> >>
> >> >> <whatever>
> >> >>
> >> >> end
> >> >>
> >> >> Nick
> >> >>
> >> >> On Mon, Aug 13, 2012 at 9:41 PM, tashi lama <ltashi32@hotmail.com> wrote:
> >> >> >
> >> >> > I wrote a do-file a while ago with 4 arguments as follows:
> >> >> >
> >> >> > begin dofile dd.do
> >> >> > clear
> >> >> > set more off
> >> >> > args graph_type broker_id start_date end_date
> >> >> > qui odbc query indigo
> >> >> > odbc load, exec("select date(read_date) as date, count(*) as hits from readership where firm_id=`broker_id' and date(read_date) between '`start_date'' and '`end_date'' group by date")
> >> >> > if "`graph_type''=="pie" {
> >> >> > ............
> >> >> > exit
> >> >> > }
> >> >> > else {
> >> >> > ........
> >> >> > exit
> >> >> > }
> >> >> >
> >> >> > The problem with this code was because args is a positional macros, whatever I enter as `1' would be graph_type, `2' as broker_id and so on. If I enter the following
> >> >> > do dd 2011-01-01 pie 2011-01-20 6429 => stata would take 2011-01-01 as graph_type and yields error.
> >> >> >
> >> >> > This makes this code not so convenient and I would like to attempt to write a code (possible ado file with synatx) so that the code would run when
> >> >> >
> >> >> > dd graph_type="pie" start_date="2011-01-01" broker_id=6429 end_date="2011-01-20"
> >> >> > or
> >> >> > dd end_date="2011-01-20" graph_type="line" start_date="2011-01-01" broker_id=6429
> >> >> >
> >> >> > I tried but got stuck
> >> >> >
> >> >> > begin adofile dd
> >> >> > clear
> >> >> > program dd
> >> >> > syntax anything
> >> >> > qui odbc ...
> >> >> > odbc load, exec(".....
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/ 		 	   		  
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/