How do the comments in the database table correspond to the fields?

[darvsum]

How do the comments in the database table correspond to the fields?
Since the meaning of comments in the database table is usually used when describing the problem, the generated SQL cannot find the field.

[zainhoda]

Could you provide an example of how you've trained it?

[darvsum]

vn.train(ddl="""
CREATE TABLE test.t_yangxin (
region_code varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'region code',
region_name varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'region name',
dept_name varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'department name',
transact_name varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'matter name',
task_name varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'task name',
transact_code varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NOT NULL COMMENT 'matter code',
transact_type varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'matter type',
transact_status varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci NULL DEFAULT NULL COMMENT 'matter status',
PRIMARY KEY (transact_code) USING BTREE
) ENGINE = InnoDB CHARACTER SET = utf8mb4 COLLATE = utf8mb4_general_ci COMMENT = 'Yangxin handling matter' ROW_FORMAT = Dynamic;
""")

vn.ask("How many pieces of data are there in t_yangxin whose matter type is public service?")

getting the wrong answer：
df = pd.read_sql_query(sql, conn)
Couldn't run sql: Execution failed on sql 'SELECT COUNT(*)
FROM t_yangxin
WHERE matter_type = 'Public Service';': (1054, "Unknown column 'matter_type' in 'where clause'")