configuration tips for new users - what worked for me

Will add, prior to timing out, I always get this error:

An error has occurred.
Please continue to the error page

An HTTP error 404 occurred. /batch?id=126&op=do

Processing continues for some time after that. I see a 404 error only once in the issues so perhapst that is a rare problem.

So it went from as described in the original comment to being able to update 50 nodes per minute consistently. Still not sure how Feeds works but here is what worked for me.

First, to create new nodes of a specific type, I think you must create a custom feed at Site Building -> Feed Importers -> New importer. When I used the standard Node import it always created a node Story type. If I did not have a content type Story I would not even realize the module was working so I am grateful I still have that content type because it caused me to continue.

Basic Settings - Use Standalone Form (not sure how Attach form works), Min refresh period (don't know if that applies to one time node creation) and Import on Submission checked Yes

Fetcher - I uploaded a csv file to my web site directory so used HTTP Fetcher. Did use both Open Office and Excel and noticed no difference.

Parser - CSV parser, default delimiter is the comma ,

Node Processor - Choose the content type to be created, I chose Filtered HTML, node never expires. Don't know how update existing node works.

Mapping - Source is the header row in the CSV file and Target is the node's CCK field name

GUID should be a unique name which is always good when importing data. I created a list number in the spreadsheet and used that.

I have a VPS so am able to choose a PHP through WHM. I tried suPHP, DSO and CGI. Don't know the difference but read there is a performance difference and ended up with CGI.

For PHP configuration, I ended up with Core Memory limit 1000 MB and unlimited - 0 - execution time and input time.

I do not know if this is optimal but it worked for me.

At first, I would start the upload, the program would seem to stall, I would get an error on the import page, and I would delete the file, create a new one, and restart the import. This lead to some duplicates.

Finally, I think I understood the program first imports all the data into its memory, and then creates the nodes. So, even if the csv file is deleted, the old file can still be in memory, causing duplicate nodes. At least, I think this is what happened.

Also, it created nodes in 50 record batches. Then it would stop and wait for cron to impor the next 50. I finally set cron at 1 minute and it seemed to work, creating 50 nodes per minute.

So, one thing I learned, if the process starts, and then there is an error on the page import page and the process seems to stall, wait a couple cron runs to see if it is still processing.

Also, prior to Feeds I had tried Node Import and Migrate and Node Import appears largely inactive / unsupported and Migrate looks great but requires the creation of a module to use it, I think. The reason I waited on Feeds though I had heard about it is I was confused by its RSS like moniker, but it works great on importing nodes.

Good luck and my impression of how Feeds works may be wrong.

Another tip:

If there is a desire to stop an import and restart it, it seems the import may continue where the last attempt stopped.

This means, if the import was stopped at record 20, the next attempt may start at 21.

By cloning the feed and giving it a new name, the problem appears to be avoided.

No, I did not have any dates to import.

Have you ever used the Rules module, you might be able to import the data as a regular cck text field and use Rules to populate the date field as the nodes are created.

Or there might be an add on module to Feeds.

Another tip:

For one time import of nodes, set: Min refresh period to never

Or I think t will try to import more in the future. May be why I had duplicates. Now that it is set to never it appears to have no more duplicates upon creation.

Another tip:

To avoid duplicates at node creation.

I think set Min refresh period to Never.

I have never tried this module except to import new data. I think it can be used to regularly check a data source but as far as updating nodes that is outside my ability. There seems to be a lot of activity with the module.