一个solr集合下是否可能具有包含集合模式的字段子集的文档？

Question

We have 4 different data sets and want to perform faceted search on them. We are currently using SolrCloud and flattened these data sets before indexing them to Solr. Even though we have relational data, our primary goal is faceted search and Solr seemed like the right option.

Rough structure of our data:

Dataset1(col1, col2, col3,col4)
Dataset2(col1,col6,col7,col8)
Dataset3(col6,col9,col10)

Flattened dataset: dataset(col1,col2,col3,col4,col6,col7,col8,col9,col10) .

In the end, we flattened them to have one common structure and have nulls where values do not exist. So far Solr works great.

Problem: Now we have additional data sets coming in and each of them have about 50-60 columns. Technically, I can still flatten these too, but I don't think it is a good idea. I know that I can have different collections with different schemas for each data set. But, we perform group by's on these documents so we need one schema.

Is there any way to maintain documents with a subset of fields of the schema under one collection without flattening them? If not, is there a better solution for this problem?

For instance:

DocA(field1, field2) DocB(field3,field4). 
Schema(field1, field2, field3, field4).

Can we have DocA and DocB under one collection with the above schema?

Our backend is on top of Cloudera Hadoop (CDH4.6 and 5.2) distribution and we can choose any tool that belongs to the Hadoop ecosystem for a possible solution.

Answer 1

Of course you can, they only need a different uniquekey for each document. If you have defined a fixed solr schema, maybe dynamicfields can help you.