[ https://issues.apache.org/jira/browse/CASSANDRA-19118?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17808755#comment-17808755 ]
Andres de la Peña commented on CASSANDRA-19118: ----------------------------------------------- [~szymon.miezal] I have just committed CASSANDRA-19245. I'll run CI once the 5.0/trunk patches are rebased. > Add support of vector type to COPY command > ------------------------------------------ > > Key: CASSANDRA-19118 > URL: https://issues.apache.org/jira/browse/CASSANDRA-19118 > Project: Cassandra > Issue Type: Bug > Components: Tool/cqlsh > Reporter: Szymon Miezal > Assignee: Szymon Miezal > Priority: Normal > Fix For: 5.0-rc, 5.x > > Time Spent: 9h 40m > Remaining Estimate: 0h > > Currently it's not possible to import rows with vector literals via {{COPY}} > command. > STR: > * Create a table > {code:sql} > CREATE TABLE testcopyfrom (id text PRIMARY KEY, embedding_vector > VECTOR<FLOAT, 6> > {code} > * Prepare csv file with sample data, for instance: > {code:sql} > 1,"[0.1, 0.2, 0.3, 0.4, 0.5, 0.6]" > 2,"[-0.1, -0.2, -0.3, -0.4, -0.5, -0.6]" {code} > * in cqlsh run > {code:sql} > COPY ks.testcopyfrom FROM data.csv > {code} > It will result in getting: > {code:sql} > TypeError: Received an argument of invalid type for column > "embedding_vector". Expected: <class 'cassandra.cqltypes.VectorType(6)'>, > Got: <class 'str'>; (required argument is not a float){code} -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org