Member since
04-12-2018
5
Posts
0
Kudos Received
0
Solutions
06-21-2018
04:41 PM
I ended up compiling the source code for the 1.6 processor and deploying that, it seems to play nicely with the 1.5 NiFi framework.
... View more
06-19-2018
11:42 AM
Thanks @Shu - upgrading to 1.6 is not an option right now so I'm implemented the DistributedMapCache workaround you linked. Edit: Seems I'd misunderstood how the PutDistributedMapCache processor works. It appears to place the flow content into the cache, but in my case the content is the spreadsheet I'm trying to convert, so replacing that with the flow attributes so I can cache them results in losing the spreadsheet in the process. The example you linked doesn't hydrate the flow content with "real" data until after the caching has taken place, so that works fine.
... View more
06-14-2018
12:07 PM
I'm using the ConvertExcelToCSVProcessor (1.5.0.3.1.0.0-564) to do what it's name describes, but it appears that my flow attributes are being dropped in the output flow. The processor correctly writes new attributes such as sheetname and numrows, but drops the existing attributes that I added earlier in the flow. Is this universal behaviour (i.e. not just my config/data), and if so is it expected? I can provide more details if it's not happening for other folks.
... View more
Labels:
- Labels:
-
Apache NiFi
04-16-2018
04:39 PM
I'm trying to build a generic NiFi flow that will take any file and run all field values through the same lookup service, regardless of how many fields are present, and what they are named. In the description for the LookupRecord processor it states: "Extracts one or more fields from a Record and looks up a value for those fields in a LookupService..." I have it working with one hard-coded field name, which obviously only looks up values in that field, but is there a way to have it look up ALL fields in my dataset against one lookup service? The description suggests I can configure to use multiple fields as the key(s), but I can't see a way to do this in a recordPath. I tried using /* as the key's recordpath, but that's not working, despite it processing all fields when used in an UpdateRecord processor. I also tried using ${field.name:prepend('/')} as the key's recordPath, and various contortions of that, but can't seem to figure out how to use this processor on all fields in the flowfile. Is there a better processor this this?
... View more
Labels:
- Labels:
-
Apache NiFi