- #Getting data from an arangodb java update
- #Getting data from an arangodb java driver
- #Getting data from an arangodb java code
I'll try to declare it in detail and I hope you can reproduce the problem. Today I've done some new tests independent from my original database. But I can query for all other attributes (without bad chars) of the bad document. The same happens if I dump the collection with "bad documents". If I query for a document with a bad character I get the UTF-8 error. I have veryfied this by searching for the key of a bad document. In most cases when I switch to a page with a bad document, the webUI will not display the document. Yes the documents are shown sometimes in the webUI. And the query mentioned above still breaks with the UTF-8 error. After a refresh of my browser window the 5 documents where no longer shown in the web interface, but the counter still says that there are 5 documents in the collection. But 5 (visible) documents could not be deleted. My last try was to delete the documents from the TestCollection one by one. I can't say why I was able to see the documensts after one import. And on both database servers (old and new) I was not able to import of your "encoding_test" documents again so that they are visible in the web interface. Also here I was not able to dump the collection with the 100 documents copied from the old database server. To make sure that no side effects from previous versions disturb my tests, I've installed arangodb 3.1.7 on a new virtual server and started my tests again. All 100 datasets where shown in the web interface but the query:Īnd I'm still not able to dump the collection.
#Getting data from an arangodb java driver
So I started to copy 100 of my corrupt data sets by a small Java program (with driver version 4.1.4) into a new collection. Everything seemed to work fine and I was able to see the imported documents in the web interface.
#Getting data from an arangodb java code
I've upgraded Arangodb on my dev system to version 3.1.7 and runned your test code again. Today I switched from Version 4.1.3 to 4.1.4 (I've done all my previous tests with v. I'm sorry, but this problem really confuses me. Is it possible that the documents are still contained in the data files preventing a successful dump? The number 4491 (under "Deletion" in row "type - Dead") seems to be the number of document with bad encoding I've deleted. It fails.īut in the info panel of M圜ollection I've seen the following:
#Getting data from an arangodb java update
I've installed the update to Version 3.1.4 and tried to dump the db again.But arangodump still fails while dumping M圜ollection with the same error message. Now I can successful query for all documents from M圜ollection. I deleted all documents with a bad encoding and also all edges referencing these documents.As I opened the new collection in the webinterface no document was shown but the documents count (lower left corner) told me that the collection was containing some documents. This also fails with an error message after a few documents where copied. I tried to copy all documents with bad encoding into a new collection.Response: 400, Error: 1501 - syntax error, unexpected end of query string near '½' at position 1:514 (while parsing) I couldn't fix that, because the query string does not contain a '½'. I tried to replace alle documents containing a bad encoding with new versions encoded to UTF-8 but several tries fail (after a lot of documents where replaced) with different error messages.It fails with the error message Invalid UTF-8 string. I tried to query for all documents from M圜ollection.Ok, I'll try to reconstruct what I've tested within the last days.
0 kommentar(er)
