We tried to get the tokenizer with Solr to work in various ways.
We found two ways that worked: posting the schema and then uploading the data, or creating the keyspace in cassandra, then posting the schema, then uploading the data.
We found that creating the keyspace in cassandra, then uploading the data, then posting the schema doesn't work.
We also found that initially posting a schema without a tokenizer and then updating the schema doesn't cause Solr to reindex the data with the tokenizer.
Is this behavior to be expected?