Word cannot find data source problem

Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but that
means I end up in the "Locate Data Source" loop with no way of opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores, that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not indicate
anything other than a fault in the display. Further, whether the Mailmerge
task pane displays the "long" data source location details or just the short
file name depends on whether the data source is in what Word thinks is the
"active" folder" (I think) - i.e. it is likely to change to that folder if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess, but in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in text
converter to read the file, or OLE DB. If it uses OLE DB, it generates a
"connection string" which contains the path name of the folder containing
the file. However, Word does not save the whole connection string (max 255
characters I think) and can truncate the pathname, so when you close and
re-open the When you close the mail merge main document and re-open it, Word
can't find the file.I'm not so sure that happens when Word opens the file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but that
means I end up in the "Locate Data Source" loop with no way of opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores, that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge files
with a user -assigned data source. One of the following will happen:
1. Our Word Merge templates (the 'letter' part of the file) are copied into
a client directory, and staff edits the file to 'link' to the new data source
(text file with CSV). Staff saves the document. The next time the staff
opens the document Word responds with 'Opening this document will run the
following SQL Command' - as we would expect - but then we get a message 'Word
cannot find its data source ' - displaying the name of the file we linked in
- which we know is there - so staff selects 'Find Data Source' and re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data source. The
only option is to remove the headers. But even after we do that, Word Merge
cannot find the data source. We cannot save the Word merge file with the
correct data source attached.

Here's the strange part. All our word files are saved to the network - the
templates and the data source. If I attached a data source that's on my
local drive to the word file, there is no problem - the merge file with the
new data source saves fine.

It's as if the merge file does not want a data source that's on a network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not indicate
anything other than a fault in the display. Further, whether the Mailmerge
task pane displays the "long" data source location details or just the short
file name depends on whether the data source is in what Word thinks is the
"active" folder" (I think) - i.e. it is likely to change to that folder if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess, but in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in text
converter to read the file, or OLE DB. If it uses OLE DB, it generates a
"connection string" which contains the path name of the folder containing
the file. However, Word does not save the whole connection string (max 255
characters I think) and can truncate the pathname, so when you close and
re-open the When you close the mail merge main document and re-open it, Word
can't find the file.I'm not so sure that happens when Word opens the file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but that
means I end up in the "Locate Data Source" loop with no way of opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores, that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

I just did some simple tests here and was able to re-open when both Mail
merge Main Document and data source were on the same network drive, i.e.
suggesting that the problem does not occur solely because it's a netwrok
drive (it worked whether I connected using E DB or the internal text file
converter).

A few questions:
a. are your mail merge document and data source in the same network folder?
b. how long is the pathname of the data source? e.g. longer or shorter than
the pathname you used when testing on a local drive? - the total length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge files
with a user -assigned data source. One of the following will happen:
1. Our Word Merge templates (the 'letter' part of the file) are copied
into
a client directory, and staff edits the file to 'link' to the new data
source
(text file with CSV). Staff saves the document. The next time the staff
opens the document Word responds with 'Opening this document will run the
following SQL Command' - as we would expect - but then we get a message
'Word
cannot find its data source ' - displaying the name of the file we linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data source.
The
only option is to remove the headers. But even after we do that, Word
Merge
cannot find the data source. We cannot save the Word merge file with the
correct data source attached.

Here's the strange part. All our word files are saved to the network -
the
templates and the data source. If I attached a data source that's on my
local drive to the word file, there is no problem - the merge file with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just the
short
file name depends on whether the data source is in what Word thinks is
the
"active" folder" (I think) - i.e. it is likely to change to that folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess, but
in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it generates a
"connection string" which contains the path name of the folder containing
the file. However, Word does not save the whole connection string (max
255
characters I think) and can truncate the pathname, so when you close and
re-open the When you close the mail merge main document and re-open it,
Word
can't find the file.I'm not so sure that happens when Word opens the file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but that
means I end up in the "Locate Data Source" loop with no way of opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores, that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both Mail
merge Main Document and data source were on the same network drive, i.e.
suggesting that the problem does not occur solely because it's a netwrok
drive (it worked whether I connected using E DB or the internal text file
converter).

A few questions:
a. are your mail merge document and data source in the same network folder?
b. how long is the pathname of the data source? e.g. longer or shorter than
the pathname you used when testing on a local drive? - the total length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge files
with a user -assigned data source. One of the following will happen:
1. Our Word Merge templates (the 'letter' part of the file) are copied
into
a client directory, and staff edits the file to 'link' to the new data
source
(text file with CSV). Staff saves the document. The next time the staff
opens the document Word responds with 'Opening this document will run the
following SQL Command' - as we would expect - but then we get a message
'Word
cannot find its data source ' - displaying the name of the file we linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data source.
The
only option is to remove the headers. But even after we do that, Word
Merge
cannot find the data source. We cannot save the Word merge file with the
correct data source attached.

Here's the strange part. All our word files are saved to the network -
the
templates and the data source. If I attached a data source that's on my
local drive to the word file, there is no problem - the merge file with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just the
short
file name depends on whether the data source is in what Word thinks is
the
"active" folder" (I think) - i.e. it is likely to change to that folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess, but
in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it generates a
"connection string" which contains the path name of the folder containing
the file. However, Word does not save the whole connection string (max
255
characters I think) and can truncate the pathname, so when you close and
re-open the When you close the mail merge main document and re-open it,
Word
can't find the file.I'm not so sure that happens when Word opens the file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but that
means I end up in the "Locate Data Source" loop with no way of opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores, that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

OK, I checked again with a similar length pathname and it does not appear to
be long enough to cause problems here. Which leaves me a bit stuck.

If you are able to test the same document and mail merge source on a much
shorter network path and it still does not work, I think that would help
establish that it is probably something to do with the network setup. I
would be looking at the permissions for the /share/ and for the underlying
folder - e.g. you may need to be able to read and write to both the share
and the folder (even though mailmerge typically only reads the data source).

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both Mail
merge Main Document and data source were on the same network drive, i.e.
suggesting that the problem does not occur solely because it's a netwrok
drive (it worked whether I connected using E DB or the internal text file
converter).

A few questions:
a. are your mail merge document and data source in the same network
folder?
b. how long is the pathname of the data source? e.g. longer or shorter
than
the pathname you used when testing on a local drive? - the total length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge
files
with a user -assigned data source. One of the following will happen:
1. Our Word Merge templates (the 'letter' part of the file) are copied
into
a client directory, and staff edits the file to 'link' to the new data
source
(text file with CSV). Staff saves the document. The next time the
staff
opens the document Word responds with 'Opening this document will run
the
following SQL Command' - as we would expect - but then we get a message
'Word
cannot find its data source ' - displaying the name of the file we
linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data source.
The
only option is to remove the headers. But even after we do that, Word
Merge
cannot find the data source. We cannot save the Word merge file with
the
correct data source attached.

Here's the strange part. All our word files are saved to the network -
the
templates and the data source. If I attached a data source that's on
my
local drive to the word file, there is no problem - the merge file with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a
network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not
indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just the
short
file name depends on whether the data source is in what Word thinks is
the
"active" folder" (I think) - i.e. it is likely to change to that
folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses
the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess,
but
in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it generates
a
"connection string" which contains the path name of the folder
containing
the file. However, Word does not save the whole connection string (max
255
characters I think) and can truncate the pathname, so when you close
and
re-open the When you close the mail merge main document and re-open
it,
Word
can't find the file.I'm not so sure that happens when Word opens the
file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname
I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that
for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data
source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and
that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but
that
means I end up in the "Locate Data Source" loop with no way of
opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in
them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But
then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file
that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores,
that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows
at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

I'll try that. I'll check into the permissions on network drives too. If
I have to I'll call Microsoft. Thanks for your help.

"Peter Jamieson" wrote:

OK, I checked again with a similar length pathname and it does not appear to
be long enough to cause problems here. Which leaves me a bit stuck.

If you are able to test the same document and mail merge source on a much
shorter network path and it still does not work, I think that would help
establish that it is probably something to do with the network setup. I
would be looking at the permissions for the /share/ and for the underlying
folder - e.g. you may need to be able to read and write to both the share
and the folder (even though mailmerge typically only reads the data source).

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both Mail
merge Main Document and data source were on the same network drive, i.e.
suggesting that the problem does not occur solely because it's a netwrok
drive (it worked whether I connected using E DB or the internal text file
converter).

A few questions:
a. are your mail merge document and data source in the same network
folder?
b. how long is the pathname of the data source? e.g. longer or shorter
than
the pathname you used when testing on a local drive? - the total length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge
files
with a user -assigned data source. One of the following will happen:
1. Our Word Merge templates (the 'letter' part of the file) are copied
into
a client directory, and staff edits the file to 'link' to the new data
source
(text file with CSV). Staff saves the document. The next time the
staff
opens the document Word responds with 'Opening this document will run
the
following SQL Command' - as we would expect - but then we get a message
'Word
cannot find its data source ' - displaying the name of the file we
linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data source.
The
only option is to remove the headers. But even after we do that, Word
Merge
cannot find the data source. We cannot save the Word merge file with
the
correct data source attached.

Here's the strange part. All our word files are saved to the network -
the
templates and the data source. If I attached a data source that's on
my
local drive to the word file, there is no problem - the merge file with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a
network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not
indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just the
short
file name depends on whether the data source is in what Word thinks is
the
"active" folder" (I think) - i.e. it is likely to change to that
folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number of
different mechanisms for opening a .tx or .csv type file and chooses
the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to guess,
but
in
Word 2002(XP) and 2003, I believe Word will either use Word's built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it generates
a
"connection string" which contains the path name of the folder
containing
the file. However, Word does not save the whole connection string (max
255
characters I think) and can truncate the pathname, so when you close
and
re-open the When you close the mail merge main document and re-open
it,
Word
can't find the file.I'm not so sure that happens when Word opens the
file
using its converter.

Anyway, if you try putting the file in a folder with a short pathname
I
think it will always work.

However, there could also be problems if Word does not recognise the
character encoding of your .csv file correctly - but let's leave that
for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its data
source for some files. It continually asks to locate the data
source
and if you point it to the file it just continually loops back and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source" and
point to the file that way.

The merge document is a one page letter. The data is a simple csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and
that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but
that
means I end up in the "Locate Data Source" loop with no way of
opening
the file at all.

The closest thing I've come up with that sort of gets me to a work
around is that some of the field headings have an underscore in
them:
First_Name, Last_Name, Address_1, etc. However, if I make a new csv
file in notepad and make the names with an _, it works fine. But
then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get the
Mail Merge side bar, step 3 has some weirdness. On the csv file
that
does not work, the source that the recipients are currently selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores,
that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it remembers
that it used to work before I changed the file name, but who knows
at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

One other thing that may be worth checking is that the document is not
connected to a template that also has a dtaa source (maybe the same one)
attached.

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
I'll try that. I'll check into the permissions on network drives too.
If
I have to I'll call Microsoft. Thanks for your help.

"Peter Jamieson" wrote:

OK, I checked again with a similar length pathname and it does not appear
to
be long enough to cause problems here. Which leaves me a bit stuck.

If you are able to test the same document and mail merge source on a much
shorter network path and it still does not work, I think that would help
establish that it is probably something to do with the network setup. I
would be looking at the permissions for the /share/ and for the
underlying
folder - e.g. you may need to be able to read and write to both the share
and the folder (even though mailmerge typically only reads the data
source).

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both
Mail
merge Main Document and data source were on the same network drive,
i.e.
suggesting that the problem does not occur solely because it's a
netwrok
drive (it worked whether I connected using E DB or the internal text
file
converter).

A few questions:
a. are your mail merge document and data source in the same network
folder?
b. how long is the pathname of the data source? e.g. longer or
shorter
than
the pathname you used when testing on a local drive? - the total
length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge
files
with a user -assigned data source. One of the following will
happen:
1. Our Word Merge templates (the 'letter' part of the file) are
copied
into
a client directory, and staff edits the file to 'link' to the new
data
source
(text file with CSV). Staff saves the document. The next time the
staff
opens the document Word responds with 'Opening this document will
run
the
following SQL Command' - as we would expect - but then we get a
message
'Word
cannot find its data source ' - displaying the name of the file we
linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data
source.
The
only option is to remove the headers. But even after we do that,
Word
Merge
cannot find the data source. We cannot save the Word merge file
with
the
correct data source attached.

Here's the strange part. All our word files are saved to the
network -
the
templates and the data source. If I attached a data source that's
on
my
local drive to the word file, there is no problem - the merge file
with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a
network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not
indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just
the
short
file name depends on whether the data source is in what Word thinks
is
the
"active" folder" (I think) - i.e. it is likely to change to that
folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number
of
different mechanisms for opening a .tx or .csv type file and
chooses
the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to
guess,
but
in
Word 2002(XP) and 2003, I believe Word will either use Word's
built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it
generates
a
"connection string" which contains the path name of the folder
containing
the file. However, Word does not save the whole connection string
(max
255
characters I think) and can truncate the pathname, so when you
close
and
re-open the When you close the mail merge main document and re-open
it,
Word
can't find the file.I'm not so sure that happens when Word opens
the
file
using its converter.

Anyway, if you try putting the file in a folder with a short
pathname
I
think it will always work.

However, there could also be problems if Word does not recognise
the
character encoding of your .csv file correctly - but let's leave
that
for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its
data
source for some files. It continually asks to locate the data
source
and if you point it to the file it just continually loops back
and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source"
and
point to the file that way.

The merge document is a one page letter. The data is a simple
csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and
that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but
that
means I end up in the "Locate Data Source" loop with no way of
opening
the file at all.

The closest thing I've come up with that sort of gets me to a
work
around is that some of the field headings have an underscore in
them:
First_Name, Last_Name, Address_1, etc. However, if I make a new
csv
file in notepad and make the names with an _, it works fine. But
then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get
the
Mail Merge side bar, step 3 has some weirdness. On the csv file
that
does not work, the source that the recipients are currently
selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores,
that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it
remembers
that it used to work before I changed the file name, but who
knows
at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

do you mean connected to a .dot file? How would I find out if it is?

"Peter Jamieson" wrote:

One other thing that may be worth checking is that the document is not
connected to a template that also has a dtaa source (maybe the same one)
attached.

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
I'll try that. I'll check into the permissions on network drives too.
If
I have to I'll call Microsoft. Thanks for your help.

"Peter Jamieson" wrote:

OK, I checked again with a similar length pathname and it does not appear
to
be long enough to cause problems here. Which leaves me a bit stuck.

If you are able to test the same document and mail merge source on a much
shorter network path and it still does not work, I think that would help
establish that it is probably something to do with the network setup. I
would be looking at the permissions for the /share/ and for the
underlying
folder - e.g. you may need to be able to read and write to both the share
and the folder (even though mailmerge typically only reads the data
source).

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both
Mail
merge Main Document and data source were on the same network drive,
i.e.
suggesting that the problem does not occur solely because it's a
netwrok
drive (it worked whether I connected using E DB or the internal text
file
converter).

A few questions:
a. are your mail merge document and data source in the same network
folder?
b. how long is the pathname of the data source? e.g. longer or
shorter
than
the pathname you used when testing on a local drive? - the total
length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word merge
files
with a user -assigned data source. One of the following will
happen:
1. Our Word Merge templates (the 'letter' part of the file) are
copied
into
a client directory, and staff edits the file to 'link' to the new
data
source
(text file with CSV). Staff saves the document. The next time the
staff
opens the document Word responds with 'Opening this document will
run
the
following SQL Command' - as we would expect - but then we get a
message
'Word
cannot find its data source ' - displaying the name of the file we
linked
in
- which we know is there - so staff selects 'Find Data Source' and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data
source.
The
only option is to remove the headers. But even after we do that,
Word
Merge
cannot find the data source. We cannot save the Word merge file
with
the
correct data source attached.

Here's the strange part. All our word files are saved to the
network -
the
templates and the data source. If I attached a data source that's
on
my
local drive to the word file, there is no problem - the merge file
with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a
network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not
indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or just
the
short
file name depends on whether the data source is in what Word thinks
is
the
"active" folder" (I think) - i.e. it is likely to change to that
folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a number
of
different mechanisms for opening a .tx or .csv type file and
chooses
the
mechanism depending on the file content (I think). I don't work for
Microsoft or have have access to the source code, so I have to
guess,
but
in
Word 2002(XP) and 2003, I believe Word will either use Word's
built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it
generates
a
"connection string" which contains the path name of the folder
containing
the file. However, Word does not save the whole connection string
(max
255
characters I think) and can truncate the pathname, so when you
close
and
re-open the When you close the mail merge main document and re-open
it,
Word
can't find the file.I'm not so sure that happens when Word opens
the
file
using its converter.

Anyway, if you try putting the file in a folder with a short
pathname
I
think it will always work.

However, there could also be problems if Word does not recognise
the
character encoding of your .csv file correctly - but let's leave
that
for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its
data
source for some files. It continually asks to locate the data
source
and if you point it to the file it just continually loops back
and
asks for the file location again. If you tell it "No" on the SQL
check dialog, the file opens and I can go to "Open Data Source"
and
point to the file that way.

The merge document is a one page letter. The data is a simple
csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder and
that
doesn't fix it.

I've tried using the registry setting to skip the SQL check, but
that
means I end up in the "Locate Data Source" loop with no way of
opening
the file at all.

The closest thing I've come up with that sort of gets me to a
work
around is that some of the field headings have an underscore in
them:
First_Name, Last_Name, Address_1, etc. However, if I make a new
csv
file in notepad and make the names with an _, it works fine. But
then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get
the
Mail Merge side bar, step 3 has some weirdness. On the csv file
that
does not work, the source that the recipients are currently
selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any underscores,
that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it
remembers
that it used to work before I changed the file name, but who
knows
at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!

Wen you have managed to open it, use Tools-Addins and Templates to have a
look.

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
do you mean connected to a .dot file? How would I find out if it is?

"Peter Jamieson" wrote:

One other thing that may be worth checking is that the document is not
connected to a template that also has a dtaa source (maybe the same one)
attached.

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
I'll try that. I'll check into the permissions on network drives
too.
If
I have to I'll call Microsoft. Thanks for your help.

"Peter Jamieson" wrote:

OK, I checked again with a similar length pathname and it does not
appear
to
be long enough to cause problems here. Which leaves me a bit stuck.

If you are able to test the same document and mail merge source on a
much
shorter network path and it still does not work, I think that would
help
establish that it is probably something to do with the network setup.
I
would be looking at the permissions for the /share/ and for the
underlying
folder - e.g. you may need to be able to read and write to both the
share
and the folder (even though mailmerge typically only reads the data
source).

--
Peter Jamieson
http://tips.pjmsn.me.uk

"Jane" wrote in message
...
1. The two files are in the same folder on a network drive.
2. the path to the data source is longish:
\\server\company\word\mhl clients\h\hollowayjane\mailmerge.txt
when I tested on the local drive it was short:
c:\cis\mailmerge.txt
3. there are no filters or sorts

"Peter Jamieson" wrote:

I just did some simple tests here and was able to re-open when both
Mail
merge Main Document and data source were on the same network drive,
i.e.
suggesting that the problem does not occur solely because it's a
netwrok
drive (it worked whether I connected using E DB or the internal
text
file
converter).

A few questions:
a. are your mail merge document and data source in the same
network
folder?
b. how long is the pathname of the data source? e.g. longer or
shorter
than
the pathname you used when testing on a local drive? - the total
length
consists of all the charaters in

\\computername\sharename\folders`filename.ext

c. have any sorts or filters been applied to the data?
--

Peter Jamieson
http://tips.pjmsn.me.uk
"Jane" wrote in message
...
I'm having a very similar problem.

In Word 2003 (on a Windows 2003 network) we cannot save word
merge
files
with a user -assigned data source. One of the following will
happen:
1. Our Word Merge templates (the 'letter' part of the file) are
copied
into
a client directory, and staff edits the file to 'link' to the new
data
source
(text file with CSV). Staff saves the document. The next time
the
staff
opens the document Word responds with 'Opening this document will
run
the
following SQL Command' - as we would expect - but then we get a
message
'Word
cannot find its data source ' - displaying the name of the file
we
linked
in
- which we know is there - so staff selects 'Find Data Source'
and
re-opens
the file. WE GET THE SAME MESSAGE - Word cannot find its data
source.
The
only option is to remove the headers. But even after we do that,
Word
Merge
cannot find the data source. We cannot save the Word merge file
with
the
correct data source attached.

Here's the strange part. All our word files are saved to the
network -
the
templates and the data source. If I attached a data source
that's
on
my
local drive to the word file, there is no problem - the merge
file
with
the
new data source saves fine.

It's as if the merge file does not want a data source that's on a
network
drive - or is it the path?


"Peter Jamieson" wrote:

Well spotted on the incorrect display of the file path name, but
unfortunately as far as I know it is a red herring and does not
indicate
anything other than a fault in the display. Further, whether the
Mailmerge
task pane displays the "long" data source location details or
just
the
short
file name depends on whether the data source is in what Word
thinks
is
the
"active" folder" (I think) - i.e. it is likely to change to that
folder
if
you create a test .csv in Word and save it.

However, things are complicated by the fact that Word has a
number
of
different mechanisms for opening a .tx or .csv type file and
chooses
the
mechanism depending on the file content (I think). I don't work
for
Microsoft or have have access to the source code, so I have to
guess,
but
in
Word 2002(XP) and 2003, I believe Word will either use Word's
built-in
text
converter to read the file, or OLE DB. If it uses OLE DB, it
generates
a
"connection string" which contains the path name of the folder
containing
the file. However, Word does not save the whole connection
string
(max
255
characters I think) and can truncate the pathname, so when you
close
and
re-open the When you close the mail merge main document and
re-open
it,
Word
can't find the file.I'm not so sure that happens when Word opens
the
file
using its converter.

Anyway, if you try putting the file in a folder with a short
pathname
I
think it will always work.

However, there could also be problems if Word does not recognise
the
character encoding of your .csv file correctly - but let's leave
that
for
now.

Peter Jamieson

wrote in message
ps.com...
Hi,

I'm having a problem getting Word to remember the link to its
data
source for some files. It continually asks to locate the data
source
and if you point it to the file it just continually loops back
and
asks for the file location again. If you tell it "No" on the
SQL
check dialog, the file opens and I can go to "Open Data
Source"
and
point to the file that way.

The merge document is a one page letter. The data is a simple
csv
file with only 10 fields.

I've tried putting the csv file in the My Data Sources folder
and
that
doesn't fix it.

I've tried using the registry setting to skip the SQL check,
but
that
means I end up in the "Locate Data Source" loop with no way of
opening
the file at all.

The closest thing I've come up with that sort of gets me to a
work
around is that some of the field headings have an underscore
in
them:
First_Name, Last_Name, Address_1, etc. However, if I make a
new
csv
file in notepad and make the names with an _, it works fine.
But
then
I get the second clue.

If I click on Tools | Letters and Mailings | Mail Merge to get
the
Mail Merge side bar, step 3 has some weirdness. On the csv
file
that
does not work, the source that the recipients are currently
selected
from appears as:

[:\foldername\source.cs] in "source.csv"

If I make a csv file from scratch and do not use any
underscores,
that
line becomes just:

"test.csv"

But as soon as I change a field name in the working csv file
to
contain and underscore it changes to:

[:\foldername\test.cs] in "test.csv"

It still works, probably because somewhere in the file it
remembers
that it used to work before I changed the file name, but who
knows
at
this point.

Has anyone else run into something similar or know of a fix?

Thanks!