Spark SQL解析报错：输入不匹配期望{<EOF>, ';'}的问题求助

你的SQL存在括号不匹配的语法错误，导致Spark解析器无法正确识别语句结构，从而抛出mismatched input 'BirthYear' expecting {, ';'}错误。

具体错误分析 #

看第一个CASE的第二个WHEN语句：

WHEN TRYCAST(CONCAT(CAST(BirthYear AS string), '-', CAST(BirthMonth AS string), '-', CAST(BirthDay AS string) AS DATE FORMAT 'YYYY-MM-DD')) IS NOT NULL THEN ...

这里CONCAT()函数的括号未闭合——你在CAST(BirthDay AS string)之后直接写了AS DATE FORMAT 'YYYY-MM-DD'，但CONCAT的参数列表还未结束（缺少闭合括号），解析器因此无法正确解析后续内容，触发语法报错。

另外还有两处细节问题：

• 同一个WHEN里的第二个CONCAT错误重复了CAST(BirthMonth AS string)，应该替换为CAST(BirthYear AS string)
• 第二个CASE语句中的TRYCAST也存在完全相同的括号不匹配问题

修复后的完整SQL #

select
    TransactionID as txn_id,
    Message as msg_txt,
    Source as src_sys,
    Firstname as indv_frst_nm,
    Middlename as indv_mdl_nm,
    Lastname as indv_last_nm,
    CASE
        WHEN BirthYear IS NULL OR BirthMonth IS NULL OR BirthDay IS NULL THEN ''
        WHEN TRYCAST(CONCAT(CAST(BirthYear AS string), '-', CAST(BirthMonth AS string), '-', CAST(BirthDay AS string)) AS DATE FORMAT 'YYYY-MM-DD') IS NOT NULL THEN 
            CAST(TRYCAST(CONCAT(CAST(BirthYear AS string), '-', CAST(BirthMonth AS string), '-', CAST(BirthDay AS string)) AS DATE FORMAT 'YYYY-MM-DD') AS string)
        ELSE '9999-12-31'
    END AS indv_brth_dt,
    DecisionCode as dcsn_cd,
    tddate as td_dt,
    RUN_ID as run_id,
    AuthorizedRepAddressCounty as auth_rep_verf_ad_cnty_nm,
    CASE
        WHEN TRYCAST(CONCAT(CAST(AuthorizedRepYearOfBirth AS string), '-', CAST(AuthorizedRepMonthOfBirth AS string), '-', CAST(AuthorizedRepDayOfBirth AS string)) AS DATE FORMAT 'YYYY-MM-DD') IS NOT NULL THEN
            CAST(TRYCAST(CONCAT(CAST(AuthorizedRepYearOfBirth AS string), '-', CAST(AuthorizedRepMonthOfBirth AS string), '-', CAST(AuthorizedRepDayOfBirth AS string)) AS DATE FORMAT 'YYYY-MM-DD') AS string)
        ELSE
            NULL
    END AS auth_rep_verf_brth_dt
from table_name;

内容的提问来源于stack exchange，提问作者INeedHelp101