Re: SQL Positional parameters: INVALID_PARAMETER_M...

RickB · ‎08-22-2024

When trying to execute a query via sql warehouse, I get the following error:
INVALID_PARAMETER_MARKER_VALUE.DUPLICATE_NAME

the sql statement uses ? placeholders and the correct number of arguments are being passed.

I am not able to use named placeholders with my library, before I refactor I am wondering if there is something I am missing.

Sample SQL:

SELECT client_ip, count(*) as requests FROM {redacted} WHERE log_event_hour >= ? AND log_event_hour <= ? AND log_event_date >= ? AND log_event_date <= ? LIMIT 10 args: [1724332222,1724335822,"2024-08-22","2024-08-22"]

szymon_dybczak · ‎08-22-2024

Hi @RickB ,

Which API are you using to invoke this? Parameter markers can be provided by:

Python using its pyspark.sql.SparkSession.sql() API.
Scala using its org.apache.spark.sql.SparkSession.sql() API.
Java using its org.apache.spark.sql.SparkSession.sql() API.
SQL

Because in each one there are subtle differences how to provide arguments to unnamed parameter markers. Take a look at below documentation entry:

Parameter markers | Databricks on AWS

If you are using sql, you should have something like below (example from documentation)

> DECLARE stmtStr = 'SELECT current_timestamp() + ?, ? * ? AS square';
> EXECUTE IMMEDIATE stmtStr USING INTERVAL '3' HOURS, 15.0, 15.0;

RickB · ‎08-23-2024

I am using the go-sql-driver: https://docs.databricks.com/en/dev-tools/go-sql-driver.html

Looking through the code it looks like it only works with named params which is a shame.

szymon_dybczak · ‎08-23-2024

Hi @RickB ,

Yeah, I've checked github repo and it looks like it currently supports only named params 😕

databricks-sql-go/parameters.go at main · databricks/databricks-sql-go · GitHub

SQL Positional parameters: INVALID_PARAMETER_MARKER_VALUE.DUPLICATE_NAME