CSV Import

[Steve Isenberg]

If the current CSV import has a simple use case then why does it always error and never import?
The error screen is no help whatsoever in determining the error.
I created a CSV with columns as displayed on the transaction input screen. Does not work.
I stripped the CSV down to the columns described below. Does not work.
What columns are supposed to be in the CSV file?
What are the formats of each column?
This is an example of the data based on the columns mentioned below:
Date,Description,Income
1/1/2017,Garri Ann Hearn,18
1/1/2017,Garri Ann Hearn,11
1/1/2017,Pamela Healy,18
1/3/2017,Robert Marten,100
1/4/2017,john landi,7
The CSV import fails consistently without any real clue as to the problem. How is this data supposed to be formatted?
Searching the archives is not helpful. There are too many unrelated results in the archive. This is a straight-forward problem and if the CSV is simple then it should not be difficult to figure this out.
Appreciate the help.Steve




      From: Geert Janssens <geert.gnucash at kobaltwit.be>
 To: Steve Isenberg <brrg58 at yahoo.com> 
 Sent: Tuesday, March 28, 2017 11:51 AM
 Subject: Re: CSV Import
   
#yiv2776656852 p, #yiv2776656852 li {white-space:pre-wrap;}On dinsdag 28 maart 2017 13:16:22 CEST you wrote:
Thanks for the info on the output columns for an export. Is there any guidance on how to do an import on split transactions? For instance does the split amounts have to be imported separately? Perhaps an example from someone who was able to import. I randomly scanned through some of the archives. Also, tried to do a search from the accounting info page using this link: Or you can search the gnucash-user Archives. https://lists.gnucash.org/search/?idxinfo=gnucash-user That results in a Not Found error for the search page. Thanks again 

Hi Steve, As I said the csv import functionality in the 2.6 series is rather limited. It has no notion of split transactions for csv data. So there is no way to import (multi-)split transactions from csv in gnucash 2.6. It will only accept data where each line describes a transaction but in e very limited way:Each line should have- a date- a description- an amount (be it deposit or withdrawal) Based on this information it will use a bayesian match algorithm to determine if the same transaction was already in your gnucash or not. On the last page of the importer it will show you its findings and allow you to correct. These corrections include* manually mark existing transactions as such if the bayesian match didn't detect it* inversely, mark transactions as new if the bayesian match falsely believed it had found a match* for all transactions that weren't matched to existing transactions in gnucash you can complete them (to a limited extent). That means for such transactions you can indicate a transfer account. (The base account was specified earlier in the import process, either in your column data or in a separate page). That's about as much as the csv importer can do. Limitations:- if your import data holds a transfer account, it's completely ignored. You will be asked to specify one in the last page (if the bayesian match didn't find a match).- there is no way to import multi-split transactions. The import will always create two-split transactions. Lastly the search pages are indeed broken. The software that provided them has been abandonded quite some time ago and doesn't work on the current server OS used to host the lists. To search the archives you can make use of google's site search by entering something like below in a google search field:<my search> site:lists.gnucash.orgSo for examplecsv import site:lists.gnucash.org Regards, Geert