Is es suitable for sparse data structure?

There is a data structure like nesting json but I can't define field name beforehand as of field name is defined by end user

I know if dynamic option is setting to true will make the fields number in mapping increase, so

1 How much is recommended maximum fields one index?
2 There are a lot of fields in one index, but there are a few not-null fields in one document in the index averagely. This condition will result in serious resource waste?

Read https://www.elastic.co/blog/sparse-versus-dense-document-values-with-apache-lucene

There is a benchmark about sparse data here :slight_smile:

https://home.apache.org/~mikemccand/lucenebench/sparseResults.html