How do I update a row without the partition key?

On my Cassandra DB with 4 Nodes, I want to execute an update statement like this.

UPDATE table_todo SET todoUserKey = '123' WHERE todoUserKey = "000";

I get an exception

InvalidRequest: Error from server: code=2200 [Ivalid query] message="Some partition key parts are missing: id"

As I understand because of Cassandra running on multiple nodes, I need to specify exactly on which node I need to execute my Update operation.

But I dont have any information about the ID. How I can perform the update statement?

You must identify the row. Cassandra requires for all write operations to provide the exact primary key. From the CQL docs of UPDATE command:
> The WHERE clause is used to select the row to update and must include
> all columns of the PRIMARY KEY

It isn't possible to execute your query because if it was allowed, it means Cassandra has to perform a full table scan to check every single row on every single partition on every single node in the cluster.

Allowing such update operate simply does not scale. If you had a table with billions of partitions, each with hundreds or thousands of rows in a cluster with hundreds of nodes, it's not difficult to see that the allowing the query to run is very expensive and will not perform well.

For this exact reason, you need to specify not just the partition key but the clustering column (where appropriate) so Cassandra can update the specific row (or rows) within the partition without having to perform a full table scan. Cheers!