pls hlp blob insert,shared server

[Anon]

I am having major problems doing a blob insert.

Seems that MySQL on my server is set up so the max_allowed_packet is 1M. I am on a shared server (cant change mysql settings) so i tried to work aroudn this by doing the following: (basically moving teh pointer down the file and inserting in segments)

$filehandle = fopen($addfile, "r");
$filesize = filesize($addfile);

$buffer = addslashes(fread($filehandle, 90000));
$query = "Insert into files (file,filename, size, userid ) values ('$buffer','$addfile_name', '$filesize')";
$this->query($query);

$id = mysql_insert_id($this->link_id());

while (!feof ($filehandle)) {
$buffer = addslashes(fread($filehandle, 90000));
$query = "UPDATE files SET file = concat(file,'$buffer') where id='$id'";
$this->query($query);
}
fclose ($filehandle);

But concat i guess is only for string fields, and not for blob fields. I have tried functions like LOAD_DATA('path/to/tmp/dir').

Does anyone know a workaround for this, this shared server @!#$ is starting to piss me the f**** off. Thanks in advance.

• FYI - for those who want to know how to change packet size etc:
  /usr/local/bin/safe_mysqld -O key_buffer=16M -O table_cache=128 -O sort_buffer=4M -O record_buffer=1M -O max_allowed_packet=24M

[turrets]

You could create a seperate table that has the folowing thre columns:
id | fileID | Fileprt

where id is an auto incriment

then in your first original table do this for the columns:

fileid | filename | size | userid

where fileid is an auto incriment.

each time you wish to update you will first insert into your orginal table the filename, size and user id. The you get thefile id it created by doing a select using that information (there may be a better way to instantly return the fileid.. but I don't know how..).
Then you will update the second table using chunks of data from the file using:
INSERT INTO filedata (fileid, filprt) VAULES ($fileid, $buffer);

When you call for the file then, you will use the file id found in the first table to locate all the chunks in the second table.. if its to be extracted, you will have to create a script to pasthrough all the chunks.

I hope this helps. it works in theory..