Dear Victor, sorry for the late response, I type the commands you told me and this is what the AMPL shows:
ampl: option tableproxy32 -v;
ampl: print _handler_desc['odbc'];
AMPL ODBC handler (20131212): expected 2-8 strings before ":[...]":
'ODBC', connection_spec ['ext_name'] [option [option...]]
Connection_spec gives a connection to the external table. If the table's
external name differs from the AMPL table name, the external name must be
given in place of 'ext_name'. For IN tables, 'SQL=sqlstmt' can appear in
place of 'ext_name', where sqlstmt is a SQL statement, such as a SELECT
statement. Possible options, explained below:
'maxlen=nnn'
'nsmix=...'
'time=...'
'verbose' or 'verbose=n' with 0 <= n <= 3
'write=append' or 'write=drop'
Use 'maxlen=nnn' to limit character strings to nnn characters (discarding
any excess characters).
With 'nsmix=*', columns of string data are treated as containing both
strings and numbers, with strings that look like decimal fixed- or floating-
point numbers treated as numbers (the default). With 'nsmix=-', columns of
string data are kept as string data. To give the effect of 'nsmix=*' to
some columns and of 'nsmix=-' to others, use syntax
index ~ 'Strcol:colname'
in the AMPL table declaration for columns that should be treated as with
the global 'nsmix=-' string. For columns to be treated in the default way,
optionally you can use syntax "index ~ 'Mixed:colname'" in the table
declaration. The "Strcol:" and "Mixed:" portion will be stripped from
the column name presented to the database. Alternatively, use 'nsmix=!'
to use the old inconsistent practice of adding quotes to strings that look
like numbers when reading but not removing the quotes when writing.
The ... in 'time=...' is a comma-separated list of names of columns that
should be regard as time data when writing tables. A better alternative
is to use syntax "index ~ 'Time:colname'" in the table declaration.
As with "Strcol:" and "Mixed:", the "Time:" portion will be stripped
from the column name shown to the database.
For OUT and INOUT tables, 'write=...' specifies how "write table" should work:
'write=drop' ==> drop and completely rewrite an existing table (default)
'write=append' ==> assume "write table" is only appending new rows
For 'verbose=n', n is the sum of
1 ==> show connection strings and
2 ==> show column types.
Plain 'verbose' is treated as 'verbose=1'.
Alternatives for connection_spec:
'filename.ext', where ext is a registered ODBC extension;
'filename.dsn' (written by the ODBC control panel's "File DSN");
an explicit connection string of the form 'DSN=...' or 'DRIVER=...';
or an ODBC data source name (see the ODBC control panel).
ampl:
I am not really sure if that is what you need but I dont know exactly how to type in the command line.
Also I am attaching the file that is giving me the problems to read the Excel files