On Tue, Aug 26, 2003 at 09:33:58AM -0700, Jeff Zucker wrote:
> Thomas A. Lowery wrote:
> know if you have other ideas that wouldn't break existing scripts.
What was I thinking ... break every CSV process in the world using
DBD::CSV. <sigh>
> DBD::CSV expects by default. You need to either add the names to the
> data file, or tell the DBD the column names for the table like so:
> $dbh->{csv_tables}->{myfile}->{col_names}=['id','name'];
This syntax presents a challenge for using dbish/DBI::Shell.
The driver configuration is done using the command line "dbi:...:..." or
DBI_DSN.
To start shell I use: dbish "dbi:CSV:f_dir=.;csv_eol=\n"
This leaves me at a dbish prompt. So to effectively use the DBD::CSV, I'll
need a plug-in to support commands like:
/csv table table_name file_name column_names
/csv table eol \n
...
Will DBD::CSV use an external configuration file or an ini like file?
Here's a test using the above connection string:
[EMAIL PROTECTED] dbish]$ dbish "dbi:CSV:f_dir=.;csv_eol=\n"
DBI::Shell 11.91 using DBI 1.38
...
Connecting to 'dbi:CSV:f_dir=.;csv_eol=\n' as 'test'...
Loaded plugins DBI::Shell::SQLMinus
[EMAIL PROTECTED]:CSV:f_dir=.;csv_eol=\n> /tab
col1,col2,col3,col4,col5
...
undef,'tlowery','testcsv','TABLE',undef
...
[17 rows of 5 fields returned]
[EMAIL PROTECTED]:CSV:f_dir=.;csv_eol=\n> select * from testcsv
[EMAIL PROTECTED]:CSV:f_dir=.;csv_eol=\n> /
nNumber,nName
'',undef
'',undef
'',undef
'gel',undef
[4 rows of 2 fields returned]
Odd, I get the column headers for no data. I need to determine where things are
getting lost.
Tom
File testcsv:
nNumber,nName
1,Tom
2,Susan
3,Phillip
4,Aaron
5,Ryan
6,Angel
7,"Little Wing"
