问SQL中的重复列名(Impala)

EN

     DROP TABLE IF EXISTS temp_nielsen_other_upc_all_markets_test;

       CREATE TABLE IF NOT EXISTS temp_nielsen_other_upc_all_markets_test
       PARTITIONED BY (all_periods_int)
       STORED AS PARQUET
       AS
        SELECT
       all_markets
       , all_periods
, dollars
FROM
(
SELECT          
                    COALESCE(gap.all_markets,fct.all_markets) AS all_markets,
                    COALESCE(gap.all_periods, fct.all_periods) AS all_periods,
                    NVL(gap.dollar,0) - NVL(fct.dollar,0) AS dollar,
                    (CAST(SUBSTR(COALESCE(gap.all_periods,fct.all_periods),7,2) AS int)+2000)*10000+CAST(SUBSTR(COALESCE(gap.all_periods,fct.all_periods),1,2) AS int)*100+CAST(SUBSTR(COALESCE(gap.all_periods,fct.all_periods),4,2) AS int) AS all_periods_int
FROM                temp_nielsen_gap_total_all_markets_raw gap
FULL OUTER JOIN(

                    SELECT
                                        all_markets,
                                        all_periods,
                                        SUM(dollar) AS dollar
                    FROM                temp_nielsen_sku_facts_all_markets_raw
                    GROUP BY            all_markets,
                                        all_periods
                ) fct
ON(                 gap.all_markets = fct.all_markets AND
                    gap.all_periods = fct.all_periods
    )
WHERE           ABS(NVL(gap.dollar,0) - NVL(fct.dollar,0)) > 1 AND
                fct.all_markets in (SELECT DISTINCT all_markets FROM temp_nielsen_gap_total_all_markets_raw)) t_final;