User can access actual columns and path variables by creating a view on a SELECT query with each required column and path variable selected from a foreign table. Each path variable must be typecast to a valid type and then aliased to the appropriate column name. This allows user to access path variables along with Parquet data. It is up to the user on what must be selected in SELECT query passed to "DataFrame.from_query()": columns, attributes, keys from foreign table and path variables.
Example
Create a view.
# While creating a view select each column is type casted to a valid type and
# then aliased to the required column name. Notice, we are selecting each attribute including path variables.
# Following is the VIEW created at the backend:
"""
REPLACE VIEW t1lview AS (
SELECT CAST($path.$var1 AS CHAR(10)) Var1,
CAST($path.$var2 AS CHAR(4)) Var2,
CAST($path.$var3 AS CHAR(2)) var3,
CAST($path.$var4 AS CHAR(2)) Var4,
a, b, c, d, e, f, g, h, i, j
FROM t1l);
"""
Create a DataFrame on the view and display the head of the DataFrame.
# Create a DataFrame on a view.
>>> wrk2dfview = DataFrame("t1lview")
>>> wrk2dfview.head().to_pandas()
Var1 Var2 var3 Var4 a b c d e f g h i j 0 csoj_files t1l. None None 10 100 110 1 CScs 1963-02-03 100.333 19.0 4 10 1 csoj_files t1l. None None 8 80 108 1 CScs 1963-02-03 200.333 19.0 4 10 2 csoj_files t1l. None None 2 20 102 1 CScs 1961-06-05 200.333 19.0 4 10 3 csoj_files t1l. None None 1 10 101 1 CScs 1961-06-05 100.333 19.0 4 10 4 csoj_files t1l. None None 9 90 109 1 CScs 1963-02-03 300.333 19.0 4 10 5 csoj_files t1l. None None 5 50 105 1 CScs 1961-06-05 200.333 19.0 4 10 6 csoj_files t1l. None None 6 60 106 1 CScs 1963-02-03 300.333 19.0 4 10 7 csoj_files t1l. None None 4 40 104 1 CScs 1961-06-05 100.333 19.0 4 10 8 csoj_files t1l. None None 3 30 103 1 CScs 1961-06-05 300.333 19.0 4 10 9 csoj_files t1l. None None 7 70 107 1 CScs 1963-02-03 100.333 19.0 4 10