[Question] How to avoid schema inference?

[hashbash]

Hello,

How to avoid schema inference when schema provided? I have complex queries and this queries consume significant resources just for schema inference (WHERE 1=0). Is it possible to avoid queries like this:

SELECT * FROM (select 7 as xxx) SPARK_GEN_SUBQ_139 WHERE 1=0

This query is generated on load():

spark = SparkSession.builder \
    .master("local[1]") \
    .appName("clickhouse-test-application") \
    .config("spark.jars.packages", "com.clickhouse:clickhouse-jdbc:0.7.1-patch1") \
    .config("spark.driver.memory", "1g") \
    .config("spark.network.timeout", "300000") \
    .config("spark.executor.heartbeatInterval", "200000") \
    .getOrCreate()


schema = StructType([
    StructField("xxx", IntegerType(), True),
])


df = (spark
    .read
    .format("jdbc")
    .schema(schema)
    .option('customSchema', 'xxx INT')
    .option('driver', 'com.clickhouse.jdbc.ClickHouseDriver')
    .option("url", "jdbc:clickhouse://host:8123")
    .option("user", 'username')
    .option("password", 'password')
    .option('query', 'select 7 as xxx')
)

df.load().collect()