How do block explorers efficiently index and query plaintext strings from OP_RETURN?

Wait 5 sec.

I saw that Blockonomics recently rolled out a native text search update that lets users query plaintext strings embedded in OP_RETURN directly from the search bar, rather than needing a specific tx hash or address first.From a utility standpoint, this is great for pulling up order IDs or data notarizations quickly. However, from a technical perspective, indexing arbitrary strings on a public ledger seems like it could get incredibly heavy.Does anyone know how block explorers typically handle the underlying architecture for plaintext string indexing?