Considering a collection in Cloud Firestore on which we do the following three queries:

1

c.where('a', '==', 'value1')
  .where('b', '==', 'value2')
  .where('c', '>=', 1)
  .orderBy('c', 'desc')

2

c.where('a', '==', 'value1')
  .where('c', '>=', 1)
  .orderBy('c', 'desc')

3

c.where('b', '==', 'value2')
  .where('c', '>=', 1)
  .orderBy('c', 'desc')

That is we order on field c, and filter on either or both of a and b.

Firestore tells me I need 3 separate indexes, one index for each of those queries, i.e.

1. a ASCENDING, b ASCENDING, c DESCENDING
2. a ASCENDING, c DESCENDING
3. b ASCENDING, c DESCENDING

Is there a way I can structure these queries or indexes so that Firestore can commonise or merge some or all of the indexes?

This is a specific illustration for clarity, but it would help anyone to understand how Firestore can and cannot combine indices.

I have all single-field indexes disabled on the collection. The key motivation is to optimise index storage size (money) on large collections (and perhaps performance).