Menu
▾
▴
Re: [mdb-dev] Success - kind of...
|
From: Andreas R. <mdb...@ri...> - 2005-10-04 01:30:35
|
Here is the promised snippet:
MdbHandle* mdb = mdb_open("<somepath>", MDB_TABLE);
mdb_read_catalog(mdb, MDB_TABLE);
MdbTableDef *table = mdb_read_table_by_name(mdb, "<sometablewithblob>",
MDB_TABLE);
mdb_read_columns(table);
mdb_read_indices(table);
mdb_rewind_table(table);
int blobLen = MDB_BIND_SIZE;
gchar blobChunk[MDB_BIND_SIZE];
int colID = mdb_bind_column_by_name(table, "<blobcolumn>", blobChunk,
&blobLen);
MdbColumn * blobCol = g_ptr_array_index(table->columns, colID - 1);
if (mdb_fetch_row(table))
{
gchar binder[MDB_MEMO_OVERHEAD];
memcpy(binder, blobChunk, MDB_MEMO_OVERHEAD);
int rlen;
if (rlen = mdb_ole_read(mdb, blobCol, binder, MDB_BIND_SIZE))
{
char* mem = malloc(rlen);
int pos = 0;
memcpy(&mem[pos], blobChunk, rlen);
pos += rlen;
while (rlen = mdb_ole_read_next(mdb, blobCol, binder))
{
mem = realloc(mem, pos + rlen);
memcpy(&mem[pos], blobChunk, rlen);
pos += rlen;
}
// Use the content of mem here.
// The length of the data is located
// in the pos variable at this point.
free(mem);
}
}
> From: Shanon Mulley <sha...@gm...>
> Reply-To: Shanon Mulley <sha...@gm...>
> Date: Tue, 4 Oct 2005 10:28:04 +1000
> To: Andreas Richter <mdb...@ri...>
> Cc: "Brian A. Seklecki" <lav...@sp...>, Sam Moffatt
> <pa...@gm...>, <mdb...@li...>
> Subject: Re: [mdb-dev] Success - kind of...
>
> So, is this to say that none of the standard mdbtools does this (like
> mdb-export, or mdb-sql) - I have to play around with the source?
>
> I'm looking forward to seeing your snippet of code, although I'm
> thinking it might break my brain getting it to work. I'm a recent
> convert to linux, and so far have avoided playing around with source
> code, but I'm willing to give it a go if need be.
>
> On 10/3/05, Andreas Richter <mdb...@ri...> wrote:
>> I can find some of the code, but don't have access from it at my current
>> location...
>> Basically mdb_bind_column to a buffer of size MDB_BIND_SIZE
>> Then after fetching the row. Copy this data into a size MDB_MEMO_OVERHEAD
>> and use the new buffer in the call to mdb_read_ole. This call will then copy
>> the data into the original bound buffer and return the size for you. Copy
>> that data into yet another buffer. Then call mdb_read_ole_next to see if
>> there is more data. If there is it will again be copied to the bound buffer.
>> You can append it to the data you already stored away. Then call
>> mdb_read_ole_next again until it returns 0. I'll paste some code here when I
>> get home tonight.
>> There is one little snippet of code in the sources that also uses
>> mdb_read_ole. That's where I figured out how to use this. Just grep for
>> mdb_read_ole in all the sources.
>> Thanks,
>> Andy
>> ----- Original Message -----
>> From: "Shanon Mulley" <sha...@gm...>
>> To: "Brian A. Seklecki" <lav...@sp...>
>> Cc: "Andreas Richter" <mdb...@ri...>; "Sam Moffatt"
>> <pa...@gm...>; <mdb...@li...>
>> Sent: Monday, October 03, 2005 8:23 AM
>> Subject: Re: [mdb-dev] Success - kind of...
>>
>>
>> Andreas,
>>
>> You seem to be having more success with reading binary data from
>> access then myself. I'm wondering if you can give me a few pointers
>> with getting binary data out. I've been teaching myself how to use
>> mdbtools, but am yet to work this out. You mention a command
>> "mdb_read_ole". Can you tell me where this is used?
>>
>> I'm using the latest version (from cvs), but didnt enable odbc - that
>> wouldnt make properly.
>>
>> Thanks.
>>
>> On 10/3/05, Brian A. Seklecki <lav...@sp...> wrote:
>>> Specifically which versions of the autotools did you use, just out of
>>> curiosity?
>>>
>>> ~BAS
>>>
>>> On Fri, 2005-09-30 at 20:59, Andreas Richter wrote:
>>>> Hi,
>>>> After beating autoconf, libtool, autogen and so on into pulp I have the
>>>> CVS
>>>> version running. It can correctly read the blobs with mdb_read_ole!!!
>>>> Yes.
>>>> However blobs that are longer than will cause an segment fault in
>>>> mdb_read_ole_next. The problem seems to be that the length calculation
>>>> is
>>>> wrong. The following code in mdb_find_pg_row comes back with a len of
>>>> 0xfffffb9b i.e. A negative number.
>>>>
>>>> mdb_swap_pgbuf(mdb);
>>>> mdb_find_row(mdb, row, off, len);
>>>> mdb_swap_pgbuf(mdb);
>>>>
>>>> In mdb_find_row the next_start is 4096 even though I am not loading the
>>>> blob
>>>> from the first row. By testing thing I think I figured out that
>>>> sometimes
>>>> blobs come with extensions, but the pg_row for the extension (i.e. The
>>>> next
>>>> pointer) is zero. Therefore making the following change in
>>>> mdb_read_ole_next
>>>> seemed to fix the crash (line 460)
>>>>
>>>> if (mdb_find_pg_row(mdb, col->cur_blob_pg_row,
>>>> &buf, &row_start, &len)) {
>>>> return 0;
>>>> }
>>>>
>>>> To
>>>>
>>>> if (!col->cur_blob_pg_row || mdb_find_pg_row(mdb,
>>>> col->cur_blob_pg_row,
>>>> &buf, &row_start, &len)) {
>>>> return 0;
>>>> }
>>>>
>>>> BTW: The CVS version of mdbtools.h has an #include <config.h> which
>>>> probably
>>>> won't work from a normal include directory.
>>>> Thanks,
>>>> Andy
>>>>
>>>>> From: Andreas Richter <mdb...@ri...>
>>>>> Date: Fri, 30 Sep 2005 11:24:31 -0400
>>>>> To: 'Sam Moffatt' <pa...@gm...>, 'Andreas Richter'
>>>>> <mdb...@ri...>
>>>>> Cc: <mdb...@li...>
>>>>> Subject: RE: [mdb-dev] Problem reading blobs
>>>>>
>>>>> No success building CVS dump under Darwin. The 0.6pre tar file did
>>>>> work
>>>>> correctly under a fink installation, but the autogen.sh file for the
>>>>> CVS
>>>>> dump did not work. Has anyone done this yet? I'll probably just
>>>>> execute the
>>>>> autogen.sh under linux and then move the files to Darwin. Any
>>>>> suggestions
>>>>> would be great.
>>>>> Thanks,
>>>>> Andy
>>>>>
>>>>> -----Original Message-----
>>>>> From: Sam Moffatt [mailto:pa...@gm...]
>>>>> Sent: Thursday, September 29, 2005 8:17 AM
>>>>> To: Andreas Richter
>>>>> Cc: mdb...@li...
>>>>> Subject: Re: [mdb-dev] Problem reading blobs
>>>>>
>>>>> what version of the library and as others will say, the cvs version
>>>>> has heaps of improvements in it.
>>>>>
>>>>> On 9/29/05, Andreas Richter <mdb...@ri...> wrote:
>>>>>>
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I am trying to read an access 2k database with blobs. After reading
>>>>>> the
>>>>> blob
>>>>>> with mdb_read_ole it will give me a 4k of data. Calling
>>>>>> mdb_read_ole_next
>>>>>> will always tell me that there is no more data. Is mdb_read_ole
>>>>>> currently
>>>>>> implemented for normal blobs? Can I help implementing it? I am using
>>>>>> the
>>>>>> library under Mac OS X.
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> Andy
>>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>> -------------------------------------------------------
>>>> This SF.Net email is sponsored by:
>>>> Power Architecture Resource Center: Free content, downloads,
>>>> discussions,
>>>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>>>> _______________________________________________
>>>> mdbtools-dev mailing list
>>>> mdb...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>>>
>>>
>>>
>>> -------------------------------------------------------
>>> This SF.Net email is sponsored by:
>>> Power Architecture Resource Center: Free content, downloads, discussions,
>>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>>> _______________________________________________
>>> mdbtools-dev mailing list
>>> mdb...@li...
>>> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by:
>> Power Architecture Resource Center: Free content, downloads, discussions,
>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>> _______________________________________________
>> mdbtools-dev mailing list
>> mdb...@li...
>> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>>
|
