Dynamic Frame writing extra columns to Redshift.

I read data from s3 using as follow.

sec_id_dyf = glueContext.create_dynamic_frame.from_options(
    connection_type = 's3',
    connection_options={'paths':['s3://<path>/sector_id_mappings.csv']},
    format = "csv",
    format_options={ "withHeader" :True}
)

Then I do necessary transformations and finally type cast to as relevant to Redshift table. Then I load these data to AWS Redshift table as follow.

from awsglue.dynamicframe import DynamicFrame

sec_id_dyf_ct = sec_id_dyf.resolveChoice(specs=[("last_letter_cell_name", "cast:string"), ("sector_id", "cast:byte")])

my_conn_opt = {
    "dbtable":"public.Q_DATA",
    "database":"dev"
}


redshift_write = glueContext.write_dynamic_frame.from_jdbc_conf(
    frame = sec_id_dyf_ct, 
    catalog_connection = "redshift-conn", 
    connection_options = my_conn_opt, 
    redshift_tmp_dir = "s3://<path2>/", 
    transformation_ctx = "redshift_write"
)

The problem is even though after all type are matched, still Glue create new column in redshift.

How to avoid this behavior in AWS Glue and Redshift. It's really appreciated if you can provide some answers for this problem. Thank you.