Sascha H.

Hi team, I'm using the docker 0.4.0 image with BigQuery. So far I was able to label and train a model. I'm facing 2 problem that I cannot seem to solve (Also looked through GH issues and gitbook docs)

(null_pointer.txt) How can I fix this? 2. After I matched the first batch of data, how do I add more data while maintaining the same z_cluster for each entitity. So existing records should keep their assigned z_cluster and new matching records should be assigned the same existing z_cluster . If I have new entries daily and want to run it through matching on a daily basis, I'd like to stitch new records to old z_cluster if they match. I tried passing a union of matched data and new data as input and added z_cluster and z_isMatch as additional columns to the matched records (and setting it to null for new records):

"fieldDefinition":[
			{
				"fieldName": "z_cluster",
				"dataType": "string",
				"fields": "z_cluster",
				"matchType": "DONT_USE"
			},
			{
				"fieldName": "z_isMatch",
				"dataType": "integer",
				"fields": "z_isMatch",
				"matchType": "DONT_USE"
			},
			{
				"fieldName": "first_user_data_point",
				"dataType": "string",
				"fields": "first_user_data_point",
				"matchType": "EXACT,NULL_OR_BLANK"
			},
			{
                         ...
                        }
]

However, when running this, I get [AMBIGUOUS_REFERENCE] Reference `z_cluster` is ambiguous, could be: [`z_cluster`, `z_cluster`]. (ambiguous_z_cluster.txt) Thank you for your help!