Hello!
I know about 4.5’s subdoc API, but the question is not about that. For example, if i have a doc like this (fields are in bytes, not chars):
{
"relatively_often_changed_field_from_10_to_1000_bytes": "0123456789"
"rarely_changed_1000_bytes_field_1": "... base64 ..."
"rarely_changed_1000_bytes_field_2": "... base64 ..."
}
does it make sense (in term of CPU / memory usage optimization for CB-servers and assuming only KV-interaction) to try to size-align such documents by introducing “padding” field and making resulting (changed) document aligned to “max possible size” ?