Adf schema mapping without an key

Question

In an copy data activity I have a hierarchical source to tabular sink mapping.

We extract data from an api.

within the data we have the following node
*,
"businessCommunication": {
"emails": [
"some-email@Company portal .com"
],
"landlines": [
""
],
"mobiles": [
""
]
},

I want to extract the email value.
In the advanced editor I tried to map this as : $['businessCommunication']['emails'][0]

But this does not work. I know I could get this to work with a DataFlow, but I'd prefer if this can stay in a CopyData activity.

Thanks

Answer 1

(EDITED, FOUND A SOLUTION)

@Palios, Jason

Given the sample data file (JSON):

{"foo":[1,2,3,4,5]}  

we can use alternative syntax to turn it into

1  
2  
3  
4  
5  

In the above picture, in the mapping I specify the collectionReference as $['foo'] and the source as foo not $['foo']. Also the type is specified in the destination.

I found the subtle difference of $[...] to be very important. When $[...] is used, this is specifying the entire path. When leaving out the $[...] it can search for anywhere in the document, including the results of cross-apply that collectionReference does.