mdbtools-dev

[mdb-dev] Using mdbtools odbc to connect to an mdb on a windows box?

[<nv...@pa...>]

Hi. I'm using a PHP script to pull data out of an access database which i=
s located on a remote windows 98SE machine. The file is on D:\shared\shar=
e\test.mdb. 

My configuration on Linux (Redhat 9.0)
odbc.ini

This is my odbc.ini

[test]
Description =3D Microsoft Access Database
Driver =3D MDBToolsODBC
Database =3D /home/nvs/test.mdb
Servername =3D 127.0.0.1
UserName =3D
Password =3D
port =3D 5432
[test2]
Description =3D Microsoft Access Database
Driver =3D MDBToolsODBC
Database =3DD:\shared\share\test.mdb
Servername =3D (ip of win machine)
UserName =3D
Password =3D
port =3D 5432


The first database works, the PHP script returns the data. However, the s=
econd one produces the following error: 
Warning: SQL error: UODB4=C2=B2=C3=B3@=C2=B8=C3=8E=C3=BF=C2=BF=E2=80=BA=C3=
=8A=C3=B1@ =E2=82=AC , SQL state 00000 in SQLConnect in /var/www/html/mdb=
.php on line 3


Looking for this problem on the net has found me a few people with the sa=
me issue, but no sollution. I am rather confused on how to set the databa=
se option for the second database. Do I use the entire path, like above, =
or do start from the shared directory? Do I need to escape backslashes? A=
ny help would be appreciated.

Greetings,
Niki Van Strydonck

RE: [mdb-dev] Segmentation Fault

[<br...@br...>]

Very well.  I squashed two bugs yesterday and have (I think) one more
to go until the patch is ready.  Hopefully this week.

Brian

On Wed, 14 Apr 2004 18:07:16 -0600, "John Finlay" wrote:

> 
> What I have is a rather old access database that is used by a
> proprietary
> WinNT program that collects data from various machines and stores it
> in the
> database. Presently, the .mdb file is about 160 megs of largely
numbers
> (counters, levels, timers, etc.) Several of these numbers are of
great
> interest to us and we've been using the (proprietary software) to
> generate a
> report each day based on the previous day's numbers and storing those
> values
> in another database I've created using PHP / MySQL. The MySQL
database
> provides much more meaningful information, and I am trying to
automate
> the
> process of extracting the information we need from the access
> database. (It
> also allows much [more] information to be gleaned from it's contents
> than
> even the reports give us so there are numerous advantages) However,
> given
> the size of the database, and the fact that each day's information is
> heaped
> on top of years of previous days' records (at about 700 rows / day),
> you can
> imagine what I wouldn't give to be able to do a "SELECT * FROM
process
> WHERE
> date = CURRENT_DATE" instead of trying to parse years of information
> for one
> day's data. :) The long and the short of it is that yes, I would be
> more
> than happy to help test the new code once it's ready for trial.
> 
> John Finlay
> IT Systems Manager
> Custom Linen Systems Ltd.
> 2925 10th Avenue N.E.
> Calgary, Alberta T2A 5L4
> Email: in...@al...
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

RE: [mdb-dev] Segmentation Fault

[Tony T. <to...@gr...>]

At 01:59 AM 2004-04-15 +0100, Fergus Allan wrote:

 >When you say table definitions, you can add upto 255 columns.  This includes
 >any deletions and additions.
<snip>
 >but any attempt to add a column gives an error in Access
 >that too many columns are defined.

Compacting the MDB will reset the "counter column" allowing you to add more 
columns.  And thus cleaning up the segmentation fault.  However a better 
long term solution would be, as Brian is doing, to update the tool to 
handle this properly.  Obviously you can't always compact at your convenience.

Tony
-----
Tony Toews, Microsoft Access MVP
Microsoft Access Links, Hints, Tips & Accounting Systems at
    http://www.granite.ab.ca/accsmstr.htm

RE: [mdb-dev] Segmentation Fault

[<br...@br...>]

Columns are never really removed.  So, you have a table with 10
columns and delete 3, and then add more rows, the new rows will have
10 columns not 7.  For fixed columns, the space used will be equal to
the length of the column (4 bytes for int's, 8 for datetime's, etc). 
For variable length columns, this will be 1 byte for JET3 or 2 bytes
for JET4.  So, not only does it use up one of the 255 "column slots"
it also uses space on disk.

Brian

On Thu, 15 Apr 2004 01:59:30 +0100, "Fergus Allan" wrote:

> 
> 
> When you say table definitions, you can add upto 255 columns.  This
> includes
> any deletions and additions.  Changes to the definitions are 'fixed'
> when
> any new data is added.
> 
> Thus
> 
> Add column
> add data
> delete column
> 
> repeated ~250 will leave you with a table with, say, 5 columns (what
> ever
> you stated with), but any attempt to add a column gives an error in
> Access
> that too many columns are defined.
> 
> 
> Regards
> 
> Fergus Allan
> 
> 
> 
> -----Original Message-----
> From: mdb...@li...
> [mailto:mdb...@li...]On Behalf Of
> br...@br...
> Sent: 15 April 2004 00:45
> To: in...@al...
> Cc: mdb...@li...
> Subject: Re: [mdb-dev] Segmentation Fault
> 
> 
> I'm currently working on these issues.  What happens is you have a
> table that has at some point had columns added or deleted from the
> table.  When this happens, Access does not go back and retouch the
> rows, it simply updates the table definition and moves on.  Thus the
> number of columns recorded for the table doesn't match the the number
> stamped on the row.  So what does this have to do with your problem?
> 
> version 0.5 largely ignored the issue and that works for a certain
> subset of column add/deletes, but can give screwy results in other
> cases (values appearing in the wrong column or corrupted for
> instance).  The CVS version is attempting to take a more correct
> approach, but isn't quite there yet.
> 
> I have a yet-to-be-checked-in patch to handle JET4 databases, but I'm
> not quite ready with JET3 (but very close).  I hope to have it all
> ready by this weekend.   If you're interested in testing, I'll let
you
> know when it's commited.
> 
> Brian
> 
> On Wed, 14 Apr 2004 15:48:37 -0600, "John Finlay" wrote:
> 
> >
> > I had successfully compiled mdbtools-0.5 and was using it to access
> > the .mdb
> > database I am working on successfully, but today when I compiled
> > mdbtools
> > from the CVS sources, it compiled 'successfully', however now every
> > time I
> > use any mdb-xxxx binaries to access the database, all display the
> same
> > result:
> >
> > I get 24 lines identical to the warning and then a segmentation
> fault.
> >
> > WARNING: number of table columns does not match number of row
> columns,
> > strange results may occur
> > Segmentation fault
> >
> > Just to verify that the mdb didn't become corrupt somehow, I tried
> the
> > 0.5
> > binaries and they still function correctly and I'm able to retrieve
> > data
> > from them. mdb-version identifies the mdb as Jet3 which is correct,
> but
> > outside of that I cannot get any useful information without
warnings
> > about
> > table columns not matching row columns and a segmentation fault.
> >
> > Has anyone else encountered this? Any ideas? System is Red Hat 9.0,
> > GLIB
> > 2.4.0, libtool-1.5, flex-2.5.4, bison-1.875, autoconf 2.59,
> > automake-1.8.3
> >
> > John Finlay
> > IT Systems Manager
> > Custom Linen Systems Ltd.
> > 2925 10th Avenue N.E.
> > Calgary, Alberta T2A 5L4
> > Email: in...@al...
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO
of
> > GenToo technologies. Learn everything from fundamentals to system
> >
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

RE: [mdb-dev] Segmentation Fault

[Fergus A. <f_w...@ya...>]

When you say table definitions, you can add upto 255 columns.  This includes
any deletions and additions.  Changes to the definitions are 'fixed' when
any new data is added.

Thus

Add column
add data
delete column

repeated ~250 will leave you with a table with, say, 5 columns (what ever
you stated with), but any attempt to add a column gives an error in Access
that too many columns are defined.


Regards

Fergus Allan



-----Original Message-----
From: mdb...@li...
[mailto:mdb...@li...]On Behalf Of
br...@br...
Sent: 15 April 2004 00:45
To: in...@al...
Cc: mdb...@li...
Subject: Re: [mdb-dev] Segmentation Fault


I'm currently working on these issues.  What happens is you have a
table that has at some point had columns added or deleted from the
table.  When this happens, Access does not go back and retouch the
rows, it simply updates the table definition and moves on.  Thus the
number of columns recorded for the table doesn't match the the number
stamped on the row.  So what does this have to do with your problem?

version 0.5 largely ignored the issue and that works for a certain
subset of column add/deletes, but can give screwy results in other
cases (values appearing in the wrong column or corrupted for
instance).  The CVS version is attempting to take a more correct
approach, but isn't quite there yet.

I have a yet-to-be-checked-in patch to handle JET4 databases, but I'm
not quite ready with JET3 (but very close).  I hope to have it all
ready by this weekend.   If you're interested in testing, I'll let you
know when it's commited.

Brian

On Wed, 14 Apr 2004 15:48:37 -0600, "John Finlay" wrote:

>
> I had successfully compiled mdbtools-0.5 and was using it to access
> the .mdb
> database I am working on successfully, but today when I compiled
> mdbtools
> from the CVS sources, it compiled 'successfully', however now every
> time I
> use any mdb-xxxx binaries to access the database, all display the
same
> result:
>
> I get 24 lines identical to the warning and then a segmentation
fault.
>
> WARNING: number of table columns does not match number of row
columns,
> strange results may occur
> Segmentation fault
>
> Just to verify that the mdb didn't become corrupt somehow, I tried
the
> 0.5
> binaries and they still function correctly and I'm able to retrieve
> data
> from them. mdb-version identifies the mdb as Jet3 which is correct,
but
> outside of that I cannot get any useful information without warnings
> about
> table columns not matching row columns and a segmentation fault.
>
> Has anyone else encountered this? Any ideas? System is Red Hat 9.0,
> GLIB
> 2.4.0, libtool-1.5, flex-2.5.4, bison-1.875, autoconf 2.59,
> automake-1.8.3
>
> John Finlay
> IT Systems Manager
> Custom Linen Systems Ltd.
> 2925 10th Avenue N.E.
> Calgary, Alberta T2A 5L4
> Email: in...@al...
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
mdbtools-dev mailing list
mdb...@li...
https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

RE: [mdb-dev] Segmentation Fault

[John F. <in...@al...>]

What I have is a rather old access database that is used by a proprietary
WinNT program that collects data from various machines and stores it in the
database. Presently, the .mdb file is about 160 megs of largely numbers
(counters, levels, timers, etc.) Several of these numbers are of great
interest to us and we've been using the (proprietary software) to generate a
report each day based on the previous day's numbers and storing those values
in another database I've created using PHP / MySQL. The MySQL database
provides much more meaningful information, and I am trying to automate the
process of extracting the information we need from the access database. (It
also allows much [more] information to be gleaned from it's contents than
even the reports give us so there are numerous advantages) However, given
the size of the database, and the fact that each day's information is heaped
on top of years of previous days' records (at about 700 rows / day), you can
imagine what I wouldn't give to be able to do a "SELECT * FROM process WHERE
date = CURRENT_DATE" instead of trying to parse years of information for one
day's data. :) The long and the short of it is that yes, I would be more
than happy to help test the new code once it's ready for trial.

John Finlay
IT Systems Manager
Custom Linen Systems Ltd.
2925 10th Avenue N.E.
Calgary, Alberta T2A 5L4
Email: in...@al...

Re: [mdb-dev] Segmentation Fault

[<br...@br...>]

I'm currently working on these issues.  What happens is you have a
table that has at some point had columns added or deleted from the
table.  When this happens, Access does not go back and retouch the
rows, it simply updates the table definition and moves on.  Thus the
number of columns recorded for the table doesn't match the the number
stamped on the row.  So what does this have to do with your problem?

version 0.5 largely ignored the issue and that works for a certain
subset of column add/deletes, but can give screwy results in other
cases (values appearing in the wrong column or corrupted for
instance).  The CVS version is attempting to take a more correct
approach, but isn't quite there yet.

I have a yet-to-be-checked-in patch to handle JET4 databases, but I'm
not quite ready with JET3 (but very close).  I hope to have it all
ready by this weekend.   If you're interested in testing, I'll let you
know when it's commited.

Brian 

On Wed, 14 Apr 2004 15:48:37 -0600, "John Finlay" wrote:

> 
> I had successfully compiled mdbtools-0.5 and was using it to access
> the .mdb
> database I am working on successfully, but today when I compiled
> mdbtools
> from the CVS sources, it compiled 'successfully', however now every
> time I
> use any mdb-xxxx binaries to access the database, all display the
same
> result:
> 
> I get 24 lines identical to the warning and then a segmentation
fault.
> 
> WARNING: number of table columns does not match number of row
columns,
> strange results may occur
> Segmentation fault
> 
> Just to verify that the mdb didn't become corrupt somehow, I tried
the
> 0.5
> binaries and they still function correctly and I'm able to retrieve
> data
> from them. mdb-version identifies the mdb as Jet3 which is correct,
but
> outside of that I cannot get any useful information without warnings
> about
> table columns not matching row columns and a segmentation fault.
> 
> Has anyone else encountered this? Any ideas? System is Red Hat 9.0,
> GLIB
> 2.4.0, libtool-1.5, flex-2.5.4, bison-1.875, autoconf 2.59,
> automake-1.8.3
> 
> John Finlay
> IT Systems Manager
> Custom Linen Systems Ltd.
> 2925 10th Avenue N.E.
> Calgary, Alberta T2A 5L4
> Email: in...@al...
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

[mdb-dev] Segmentation Fault

[John F. <in...@al...>]

I had successfully compiled mdbtools-0.5 and was using it to access the .mdb
database I am working on successfully, but today when I compiled mdbtools
from the CVS sources, it compiled 'successfully', however now every time I
use any mdb-xxxx binaries to access the database, all display the same
result:

I get 24 lines identical to the warning and then a segmentation fault.

WARNING: number of table columns does not match number of row columns,
strange results may occur
Segmentation fault

Just to verify that the mdb didn't become corrupt somehow, I tried the 0.5
binaries and they still function correctly and I'm able to retrieve data
from them. mdb-version identifies the mdb as Jet3 which is correct, but
outside of that I cannot get any useful information without warnings about
table columns not matching row columns and a segmentation fault.

Has anyone else encountered this? Any ideas? System is Red Hat 9.0, GLIB
2.4.0, libtool-1.5, flex-2.5.4, bison-1.875, autoconf 2.59, automake-1.8.3

John Finlay
IT Systems Manager
Custom Linen Systems Ltd.
2925 10th Avenue N.E.
Calgary, Alberta T2A 5L4
Email: in...@al...

[mdb-dev] (no subject)

[John F. <in...@al...>]

I had successfully compiled mdbtools-0.5 and was using it to access the .mdb
database I am working on successfully, but today when I compiled mdbtools
from the CVS sources, it compiled 'successfully', however now every time I
use any mdb-xxxx binaries to access the database, all display the same
result:

I get 24 lines identical to the warning and then a segmentation fault.

WARNING: number of table columns does not match number of row columns,
strange results may occur
Segmentation fault

Just to verify that the mdb didn't become corrupt somehow, I tried the 0.5
binaries and they still function correctly and I'm able to retrieve data
from them. mdb-version identifies the mdb as Jet3 which is correct, but
outside of that I cannot get any useful information without warnings about
table columns not matching row columns and a segmentation fault.

Has anyone else encountered this? Any ideas? System is Red Hat 9.0, GLIB
2.4.0, libtool-1.5, flex-2.5.4, bison-1.875, autoconf 2.59, automake-1.8.3

John Finlay
IT Systems Manager
Custom Linen Systems Ltd.
2925 10th Avenue N.E.
Calgary, Alberta T2A 5L4
Email: in...@al...

[mdb-dev] Re: mdbtools-dev digest, Vol 1 #363 - 2 msgs

[michael <mi...@tt...>]

Jeff wrote:

>   What are you using for the JDBC-ODBC bridge, and how is it set up?
> I am afraid I don't know much about JDBC, but unixODBC must be
> working or isql would not have worked.  (Sorry for taking so long to
> respond.)

Hi, thanks for replying! 

I'm using the stock Sun JDBC-ODBC bridge driver which comes with the Sun 
j2sdk, and unixODBC (2.2.3-6 from the Redhat 9.0 RPM), both of which 
seem to work fine with the postgres ODBC driver that is packaged in that
RPM of UnixODBC.

As for my setup, I just followed pretty much to the letter the information 
on Bryan Mills' webpage: http://bryanmills.net:8086/archives/000099.html

Thanks for your time!
Michael

> > 
> > 
> > Hi.. As part of the course I'm currently studying, we are asked to use 
> > Java to manipulate an MS Access database. 
> > 
> > The general idea is that we do this using Java, and the JDBC-ODBC 
> > gateway for Windows. However, the course team have said that they would 
> > be interested to hear of success from those of us who are using Linux. 
> > 
> > This is how I came to be experimenting with mdbtools.
> > 
> > I can access an mdb database file using the various stand-alone utilities 
> > that come as part of the mdbtools package, but so far,  I'm having no 
> > success interfacing the database to the ODBC framework using the 
> > libmdbodbc library that comes as part of mdbtools and consequently my 
> > attempts at accessing the database via the JDBC-ODBC bridge are also 
> > failing. 
> > 
> > A guy called Bryan Mills has a webpage documenting his success using  
> > the mdb ODBC driver and UnixODBC (which he accesses with PHP rather than 
> > Java) but his success seems to be unique. 
> > 
> > I can't find anyone else on the net who has duplicated it. Bryan's 
> > webpage is at 
> > 
> > I've followed Bryan Mills' example for creating the UnixODBC template 
> > files (where my odbcinst.ini references the mdb driver library and 
> > odbc.ini references the DSN and its database)  and then I've used the 
> > isql tool which comes as part of UnixODBC to successfully query the mdb 
> > database, but that's about as far as my successes have gone.
> > 
> > 	root@home mdbtools]# isql M360products
> > 	+---------------------------------------+
> > 	| Connected!                            |
> > 	| sql-statement                         | 
> > 	| help [tablename]                      |
> > 	| quit                                  |                                       
> > 	+---------------------------------------+
> > 	SQL> select ProductName,Quantity from StoreProducts where Quantity<50
> > 	+---------------------------------------------------+---------+
> > 	| ProductName                                       | Quantity|
> > 	+---------------------------------------------------+---------+
> > 	| Blue Shirts                                       | 30      |
> > 	| Green Shirts                                      | 0       |
> > 	| Grey Trousers                                     | 0       |
> > 	| Khaki Trousers                                    | 33      |
> > 	| White Jackets                                     | 22      |
> > 	| Green Jackets                                     | 44      |
> > 	+---------------------------------------------------+---------+
> > 	6 rows affected
> > 	6 rows returned
> > 	SQL>
> > 
> > The following show the contents of my odbc.ini file, which should, as far 
> > as I understand it, bind the MDB ODBC service to port 5432. I'm curious 
> > to find that netstat doesn't show anything listening on 5432. Is that 
> > an indication that something is wrong?
> > 
> > 	[root@home lib]# cat /etc/odbc.ini
> > 
> > 	[ODBC Data Sources]
> > 	M360products = libmdbodbc
> > 
> > 	[M360products]
> > 	Driver		= /usr/local/lib/libmdbodbc.so
> > 	Description	= M360 products
> > 	Database	= /home/michael/M360/TextExercises/C5Code/5.1/src/Ex51/ProductsExercise.mdb
> > 	ServerName	= localhost
> > 	Port		= 5432
> > 
> > Using the odbcinst tool to query UnixODBC shows that the MS Access driver 
> > has apparently been successfully loaded by UnixODBC, and that my demo 
> > MS Access database, ProductExercise.mdh has been successfully registered 
> > on the ODBC resource with the Data Source Name  M360Products.
> > 
> > 	[root@home lib]# odbcinst -q -d
> > 	[MDBToolsODBC]
> > 
> > 	[root@ttfn35 lib]# odbcinst -q -s
> > 	[M360products]
> > 	[ODBC]
> > 
> > Unfortunately, when I execute Java code known to work for Windows ODBC
> > and which should issue an SQL query against this database through the 
> > JDBC-ODBC gateway, I get the following error reported:
> > 
> > 	[root@home Ex51]# java ex51/QuerySolution
> > 	***Solution being executed***...
> > 	Driver loaded
> > 	Problem accessing the database java.sql.SQLException: 
> > 	[unixODBC][Driver Manager]Data source name not found, and no default driver specified
> > 
> > I've tried using the mdbodbc driver that comes with the latest version of 
> > mdbtools packaged in RPM format, and I've built mdbtools from the source 
> > tar ball on the mdbtools site and I've tried the latest version from CVS 
> > (which apparently now supports mdb write) but sadly all have stalled at 
> > this same hurdle. 
> > 
> > I'm currently moving from UnixODBC to iODBC in order to rule out that as 
> > the problem.
> > 
> > Would someone kind here please point out what I'm doing wrong, and how 
> > they got the ODBD mdb driver to function as it is intended.
> > 
> > Thank you (and apologies for the long-winded posting!)
> > 
> > Michael
> 
> 
> 	
> 		
> __________________________________
> Do you Yahoo!?
> Yahoo! Small Business $15K Web Design Giveaway 
> http://promotions.yahoo.com/design_giveaway/
> 
> 
> 
> --__--__--
> 
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> 
> 
> End of mdbtools-dev Digest
> 

[mdb-dev] Compiling out of the box

[Jeff S. <why...@ya...>]

FWIW, I am excited to finally have this sequence working for me with no intervention:

  cvs co mdbtools
  cd mdbtools
  ./autogen.sh --prefix=$HOME/odbc --with-unixodbc=$HOME/odbc
  make
  make install

My change to configure.in helped --with-unixodbc work fully, and updating acinclude.m4 to
the latest libtool.m4 kept me from having to hand-edit the generated libtool every time
that autogen.sh was run.

 -- Jeff Smith


	
		
__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

Re: [mdb-dev] MDBTools and UnixODBC

[Jeff S. <why...@ya...>]

Michael,
  What are you using for the JDBC-ODBC bridge, and how is it set up?
I am afraid I don't know much about JDBC, but unixODBC must be
working or isql would not have worked.  (Sorry for taking so long to
respond.)

 -- Jeff Smith


--- michael <mi...@tt...> wrote:
> 
> 
> Hi.. As part of the course I'm currently studying, we are asked to use 
> Java to manipulate an MS Access database. 
> 
> The general idea is that we do this using Java, and the JDBC-ODBC 
> gateway for Windows. However, the course team have said that they would 
> be interested to hear of success from those of us who are using Linux. 
> 
> This is how I came to be experimenting with mdbtools.
> 
> I can access an mdb database file using the various stand-alone utilities 
> that come as part of the mdbtools package, but so far,  I'm having no 
> success interfacing the database to the ODBC framework using the 
> libmdbodbc library that comes as part of mdbtools and consequently my 
> attempts at accessing the database via the JDBC-ODBC bridge are also 
> failing. 
> 
> A guy called Bryan Mills has a webpage documenting his success using  
> the mdb ODBC driver and UnixODBC (which he accesses with PHP rather than 
> Java) but his success seems to be unique. 
> 
> I can't find anyone else on the net who has duplicated it. Bryan's 
> webpage is at http://bryanmills.net:8086/archives/000099.html
> 
> I've followed Bryan Mills' example for creating the UnixODBC template 
> files (where my odbcinst.ini references the mdb driver library and 
> odbc.ini references the DSN and its database)  and then I've used the 
> isql tool which comes as part of UnixODBC to successfully query the mdb 
> database, but that's about as far as my successes have gone.
> 
> 	root@home mdbtools]# isql M360products
> 	+---------------------------------------+
> 	| Connected!                            |
> 	| sql-statement                         | 
> 	| help [tablename]                      |
> 	| quit                                  |                                       
> 	+---------------------------------------+
> 	SQL> select ProductName,Quantity from StoreProducts where Quantity<50
> 	+---------------------------------------------------+---------+
> 	| ProductName                                       | Quantity|
> 	+---------------------------------------------------+---------+
> 	| Blue Shirts                                       | 30      |
> 	| Green Shirts                                      | 0       |
> 	| Grey Trousers                                     | 0       |
> 	| Khaki Trousers                                    | 33      |
> 	| White Jackets                                     | 22      |
> 	| Green Jackets                                     | 44      |
> 	+---------------------------------------------------+---------+
> 	6 rows affected
> 	6 rows returned
> 	SQL>
> 
> The following show the contents of my odbc.ini file, which should, as far 
> as I understand it, bind the MDB ODBC service to port 5432. I'm curious 
> to find that netstat doesn't show anything listening on 5432. Is that 
> an indication that something is wrong?
> 
> 	[root@home lib]# cat /etc/odbc.ini
> 
> 	[ODBC Data Sources]
> 	M360products = libmdbodbc
> 
> 	[M360products]
> 	Driver		= /usr/local/lib/libmdbodbc.so
> 	Description	= M360 products
> 	Database	= /home/michael/M360/TextExercises/C5Code/5.1/src/Ex51/ProductsExercise.mdb
> 	ServerName	= localhost
> 	Port		= 5432
> 
> Using the odbcinst tool to query UnixODBC shows that the MS Access driver 
> has apparently been successfully loaded by UnixODBC, and that my demo 
> MS Access database, ProductExercise.mdh has been successfully registered 
> on the ODBC resource with the Data Source Name  M360Products.
> 
> 	[root@home lib]# odbcinst -q -d
> 	[MDBToolsODBC]
> 
> 	[root@ttfn35 lib]# odbcinst -q -s
> 	[M360products]
> 	[ODBC]
> 
> Unfortunately, when I execute Java code known to work for Windows ODBC
> and which should issue an SQL query against this database through the 
> JDBC-ODBC gateway, I get the following error reported:
> 
> 	[root@home Ex51]# java ex51/QuerySolution
> 	***Solution being executed***...
> 	Driver loaded
> 	Problem accessing the database java.sql.SQLException: 
> 	[unixODBC][Driver Manager]Data source name not found, and no default driver specified
> 
> I've tried using the mdbodbc driver that comes with the latest version of 
> mdbtools packaged in RPM format, and I've built mdbtools from the source 
> tar ball on the mdbtools site and I've tried the latest version from CVS 
> (which apparently now supports mdb write) but sadly all have stalled at 
> this same hurdle. 
> 
> I'm currently moving from UnixODBC to iODBC in order to rule out that as 
> the problem.
> 
> Would someone kind here please point out what I'm doing wrong, and how 
> they got the ODBD mdb driver to function as it is intended.
> 
> Thank you (and apologies for the long-winded posting!)
> 
> Michael


	
		
__________________________________
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway 
http://promotions.yahoo.com/design_giveaway/

[mdb-dev] mdb_close

[Jeff S. <why...@ya...>]

For mdb_close to parallel mdb_open, it needs to call mdb_free_handle at the end.  As it
stands, calling mdb_close without mdb_free_handle, or vice-a-versa, causes memory leaks. 
  Most of the utilities only call mdb_free_handle.

I can make this change, and all the changes to utilities and such to use it properly. 
But as this would be a change in the interface behavior, I figured it should be cleared
here first.

 -- Jeff Smith

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[mdb-dev] libmdb reference manual

[<br...@br...>]

Hi all,

Well I spent tonight getting to know the gtk-doc tools so we could
generate a proper reference manual for the libraries.  It's still in
the  preliminary stages but if you look in doc/reference/libmdb,
you'll see the beginnings.  To get it to work you'll need the gtk-doc
tools on your system.  The configure script *should* detect them and
build or not build accordingly, if this doesn't work for you
(especially erroring when the tools are not present) please let me
know.

There are only a handful of documented functions at present, but
that'll get better soon.   Once it's in some kind of order, I'll post
it on the web page.

Note on 0.6pre1 status:  I have JET4 tables with deleted/added columns 
working to my satisfaction and am working on JET3. Once that's done
I'll be rolling up 0.6pre1 and then heading on to work on more of the
write support.

Brian

Re: [mdb-dev] Memory leak fix

[<br...@br...>]

I just run `date` when I start writing the ChangeLog... emacs has some
super-special mode to edit the thing, but alas I always have and
probably always will be a vi guy.

Brian

On Mon, 12 Apr 2004 12:39:07 -0700 (PDT), Jeff Smith wrote:

> 
> Brian,
>   Thanks, I'll do my best to be a good citizen. :-)
> 
>   One question on the ChangeLog format: where do you pull the date
> from?  i.e. Mon Apr 12
> 09:56:23 EDT 2004 for the latest commit.  Do you just run 'date' and
> paste it in?  Or is
> there some more complex system you use?  Does it really matter? :-p
> 
>  -- Jeff Smith
> 
> --- br...@br... wrote:
> > Ok, you're all set up.  Ummm, ground rules (nothing too onerous)
> > 
> > 1. Keep the ChangeLog current.  If you checked it in, it should
have
> a
> > ChangeLog line.
> > 
> > 2. API additions/changes in libmdb should be posted to the list
first
> > for comments.  For one liners and simple bug fixes, just commit it.
> > 
> > 3. For ODBC stuff, just post a summary of what you did to the list
> > after commiting (for big stuff).  Since the API is externally
> defined,
> > there's not much to worry about as far as those decisions go.  The
> > driver currently only works with unixODBC, but I'd like to leave
the
> > door open to fixing the iODBC problems, so please ifdef any
> > unixODBC'isms.
> > 
> > That's about it.  Style as you have seen is basically K&R with the
> > exception of return types on the preceeding line from the function
> > name. Generally, I'm not a big fan of the GNU style.  Variables are
> > lower_case not mixedCase or MixedCase.  Tabs, not spaces.  Other
> > questions on style, just ask.
> > 
> > Brian
> > 
> > On Mon, 12 Apr 2004 09:12:54 -0700 (PDT), Jeff Smith wrote:
> > 
> > > 
> > > Brian,
> > >   Yes, I will accept your gracious offer.  :-)
> > > I have several things in the works in addition to what I have
sent
> to
> > > the list, so commit access would make the process a lot smoother.
> > > 
> > >  -- Jeff Smith
> > > 
> > > 
> > > --- br...@br... wrote:
> > > > Jeff,
> > > > 
> > > > I commited this but haven't gotten to the other 4 just yet,
today
> > > > probably.  You seem to be in it for the long haul, do you want
> > commit
> > > > rights?
> > > > 
> > > > Brian
> > > > 
> > > > On Sun, 11 Apr 2004 21:51:29 -0700 (PDT), Jeff Smith wrote:
> > > > 
> > > > > 
> > > > > fixing a memory leak in mdb_exit
> 
> __________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online by April 15th
> http://taxes.yahoo.com/filing.html

Re: [mdb-dev] Memory leak fix

[Jeff S. <why...@ya...>]

Brian,
  Thanks, I'll do my best to be a good citizen. :-)

  One question on the ChangeLog format: where do you pull the date from?  i.e. Mon Apr 12
09:56:23 EDT 2004 for the latest commit.  Do you just run 'date' and paste it in?  Or is
there some more complex system you use?  Does it really matter? :-p

 -- Jeff Smith

--- br...@br... wrote:
> Ok, you're all set up.  Ummm, ground rules (nothing too onerous)
> 
> 1. Keep the ChangeLog current.  If you checked it in, it should have a
> ChangeLog line.
> 
> 2. API additions/changes in libmdb should be posted to the list first
> for comments.  For one liners and simple bug fixes, just commit it.
> 
> 3. For ODBC stuff, just post a summary of what you did to the list
> after commiting (for big stuff).  Since the API is externally defined,
> there's not much to worry about as far as those decisions go.  The
> driver currently only works with unixODBC, but I'd like to leave the
> door open to fixing the iODBC problems, so please ifdef any
> unixODBC'isms.
> 
> That's about it.  Style as you have seen is basically K&R with the
> exception of return types on the preceeding line from the function
> name. Generally, I'm not a big fan of the GNU style.  Variables are
> lower_case not mixedCase or MixedCase.  Tabs, not spaces.  Other
> questions on style, just ask.
> 
> Brian
> 
> On Mon, 12 Apr 2004 09:12:54 -0700 (PDT), Jeff Smith wrote:
> 
> > 
> > Brian,
> >   Yes, I will accept your gracious offer.  :-)
> > I have several things in the works in addition to what I have sent to
> > the list, so commit access would make the process a lot smoother.
> > 
> >  -- Jeff Smith
> > 
> > 
> > --- br...@br... wrote:
> > > Jeff,
> > > 
> > > I commited this but haven't gotten to the other 4 just yet, today
> > > probably.  You seem to be in it for the long haul, do you want
> commit
> > > rights?
> > > 
> > > Brian
> > > 
> > > On Sun, 11 Apr 2004 21:51:29 -0700 (PDT), Jeff Smith wrote:
> > > 
> > > > 
> > > > fixing a memory leak in mdb_exit

__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

Re: [mdb-dev] Memory leak fix

[<br...@br...>]

Ok, you're all set up.  Ummm, ground rules (nothing too onerous)

1. Keep the ChangeLog current.  If you checked it in, it should have a
ChangeLog line.

2. API additions/changes in libmdb should be posted to the list first
for comments.  For one liners and simple bug fixes, just commit it.

3. For ODBC stuff, just post a summary of what you did to the list
after commiting (for big stuff).  Since the API is externally defined,
there's not much to worry about as far as those decisions go.  The
driver currently only works with unixODBC, but I'd like to leave the
door open to fixing the iODBC problems, so please ifdef any
unixODBC'isms.

That's about it.  Style as you have seen is basically K&R with the
exception of return types on the preceeding line from the function
name. Generally, I'm not a big fan of the GNU style.  Variables are
lower_case not mixedCase or MixedCase.  Tabs, not spaces.  Other
questions on style, just ask.

Brian

On Mon, 12 Apr 2004 09:12:54 -0700 (PDT), Jeff Smith wrote:

> 
> Brian,
>   Yes, I will accept your gracious offer.  :-)
> I have several things in the works in addition to what I have sent to
> the list, so commit access would make the process a lot smoother.
> 
>  -- Jeff Smith
> 
> 
> --- br...@br... wrote:
> > Jeff,
> > 
> > I commited this but haven't gotten to the other 4 just yet, today
> > probably.  You seem to be in it for the long haul, do you want
commit
> > rights?
> > 
> > Brian
> > 
> > On Sun, 11 Apr 2004 21:51:29 -0700 (PDT), Jeff Smith wrote:
> > 
> > > 
> > > fixing a memory leak in mdb_exit
> 
> __________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online by April 15th
> http://taxes.yahoo.com/filing.html
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

Re: [mdb-dev] Memory leak fix

[Jeff S. <why...@ya...>]

Brian,
  Yes, I will accept your gracious offer.  :-)
I have several things in the works in addition to what I have sent to
the list, so commit access would make the process a lot smoother.

 -- Jeff Smith


--- br...@br... wrote:
> Jeff,
> 
> I commited this but haven't gotten to the other 4 just yet, today
> probably.  You seem to be in it for the long haul, do you want commit
> rights?
> 
> Brian
> 
> On Sun, 11 Apr 2004 21:51:29 -0700 (PDT), Jeff Smith wrote:
> 
> > 
> > fixing a memory leak in mdb_exit

__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

Re: [mdb-dev] Memory leak fix

[<br...@br...>]

Jeff,

I commited this but haven't gotten to the other 4 just yet, today
probably.  You seem to be in it for the long haul, do you want commit
rights?

Brian

On Sun, 11 Apr 2004 21:51:29 -0700 (PDT), Jeff Smith wrote:

> 
> fixing a memory leak in mdb_exit
> 
> __________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online by April 15th
> http://taxes.yahoo.com/filing.html

[mdb-dev] More cruft

[Jeff S. <why...@ya...>]

The only config.h.in being utilized is in include/.  Please remove all others:

./config.h.in
./src/extras/config.h.in
./src/config.h.in
./src/include/config.h.in -- Of course, you are deleting this whole directory, right? :-)
./src/libmdb/config.h.in
./src/util/config.h.in

 -- Jeff Smith

__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

[mdb-dev] Memory leak fix

[Jeff S. <why...@ya...>]

fixing a memory leak in mdb_exit

__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

Re: [mdb-dev] src/include directory

[Leonard L. <lle...@ma...>]

This is a very good point Jeff.  In fact, I was using the mdbtools.h
file in that directory for most of my development until I found the
other one.

--
Lenny

On Sat, 2004-04-10 at 17:14, Jeff Smith wrote:
> I think everything in src/include has been moved elsewhere.  Brian, can you please remove
> this directory (to the extent possible with cvs) and everything in it?  Leaving stuff in
> multiple places can only lead to confusion.
> 
>  -- Thanks,
>  -- Jeff Smith
> 
> __________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online by April 15th
> http://taxes.yahoo.com/filing.html
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> 

[mdb-dev] MDBTools and UnixODBC

[michael <mi...@tt...>]

Hi.. As part of the course I'm currently studying, we are asked to use 
Java to manipulate an MS Access database. 

The general idea is that we do this using Java, and the JDBC-ODBC 
gateway for Windows. However, the course team have said that they would 
be interested to hear of success from those of us who are using Linux. 

This is how I came to be experimenting with mdbtools.

I can access an mdb database file using the various stand-alone utilities 
that come as part of the mdbtools package, but so far,  I'm having no 
success interfacing the database to the ODBC framework using the 
libmdbodbc library that comes as part of mdbtools and consequently my 
attempts at accessing the database via the JDBC-ODBC bridge are also 
failing. 

A guy called Bryan Mills has a webpage documenting his success using  
the mdb ODBC driver and UnixODBC (which he accesses with PHP rather than 
Java) but his success seems to be unique. 

I can't find anyone else on the net who has duplicated it. Bryan's 
webpage is at http://bryanmills.net:8086/archives/000099.html

I've followed Bryan Mills' example for creating the UnixODBC template 
files (where my odbcinst.ini references the mdb driver library and 
odbc.ini references the DSN and its database)  and then I've used the 
isql tool which comes as part of UnixODBC to successfully query the mdb 
database, but that's about as far as my successes have gone.

	root@home mdbtools]# isql M360products
	+---------------------------------------+
	| Connected!                            |
	| sql-statement                         | 
	| help [tablename]                      |
	| quit                                  |                                       
	+---------------------------------------+
	SQL> select ProductName,Quantity from StoreProducts where Quantity<50
	+---------------------------------------------------+---------+
	| ProductName                                       | Quantity|
	+---------------------------------------------------+---------+
	| Blue Shirts                                       | 30      |
	| Green Shirts                                      | 0       |
	| Grey Trousers                                     | 0       |
	| Khaki Trousers                                    | 33      |
	| White Jackets                                     | 22      |
	| Green Jackets                                     | 44      |
	+---------------------------------------------------+---------+
	6 rows affected
	6 rows returned
	SQL>

The following show the contents of my odbc.ini file, which should, as far 
as I understand it, bind the MDB ODBC service to port 5432. I'm curious 
to find that netstat doesn't show anything listening on 5432. Is that 
an indication that something is wrong?

	[root@home lib]# cat /etc/odbc.ini

	[ODBC Data Sources]
	M360products = libmdbodbc

	[M360products]
	Driver		= /usr/local/lib/libmdbodbc.so
	Description	= M360 products
	Database	= /home/michael/M360/TextExercises/C5Code/5.1/src/Ex51/ProductsExercise.mdb
	ServerName	= localhost
	Port		= 5432

Using the odbcinst tool to query UnixODBC shows that the MS Access driver 
has apparently been successfully loaded by UnixODBC, and that my demo 
MS Access database, ProductExercise.mdh has been successfully registered 
on the ODBC resource with the Data Source Name  M360Products.

	[root@home lib]# odbcinst -q -d
	[MDBToolsODBC]

	[root@ttfn35 lib]# odbcinst -q -s
	[M360products]
	[ODBC]

Unfortunately, when I execute Java code known to work for Windows ODBC
and which should issue an SQL query against this database through the 
JDBC-ODBC gateway, I get the following error reported:

	[root@home Ex51]# java ex51/QuerySolution
	***Solution being executed***...
	Driver loaded
	Problem accessing the database java.sql.SQLException: 
	[unixODBC][Driver Manager]Data source name not found, and no default driver specified

I've tried using the mdbodbc driver that comes with the latest version of 
mdbtools packaged in RPM format, and I've built mdbtools from the source 
tar ball on the mdbtools site and I've tried the latest version from CVS 
(which apparently now supports mdb write) but sadly all have stalled at 
this same hurdle. 

I'm currently moving from UnixODBC to iODBC in order to rule out that as 
the problem.

Would someone kind here please point out what I'm doing wrong, and how 
they got the ODBD mdb driver to function as it is intended.

Thank you (and apologies for the long-winded posting!)

Michael

[mdb-dev] src/include directory

[Jeff S. <why...@ya...>]

I think everything in src/include has been moved elsewhere.  Brian, can you please remove
this directory (to the extent possible with cvs) and everything in it?  Leaving stuff in
multiple places can only lead to confusion.

 -- Thanks,
 -- Jeff Smith

__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html

[mdb-dev] random patches 4/4

[Jeff S. <why...@ya...>]

src/odbc/odbc.c:
  utilize _odbc_get_string_size in SQLPrepare

__________________________________
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway 
http://promotions.yahoo.com/design_giveaway/