使用sql查找具有不同拼写（如缩写）的相同地址

Question

我想找到具有不同拼写/表达（缩写）示例的相同地址，如下所示。但我的查询不起作用。例如 St-->Street（正确）但 Street-->Streetict。 这是代码。

SELECT B_ADDRESS,H_ADDRESS, REGEXP_REPLACE(REGEXP_REPLACE(B_ADDRESS,'S[a-z]r.|(St.)','Street'),'(Dist.)|(Dt.)','Street') as B_NEW,
        REGEXP_REPLACE(REGEXP_REPLACE(H_ADDRESS,'(Str.)|(St.)','Street'), '(Dist.)|(Dt.)','District') as H_NEW
        FROM (

 SELECT H_ADDRESS,B_ADDRESS
                FROM
                ( 
                    SELECT 'Washington Str. No:60 ABD' AS H_ADDRESS,'Washington Street No60 ABD' AS B_ADDRESS FROM DUAL UNION ALL 
                    SELECT 'Pennsylvania Dt. St. No 6 ABD' AS H_ADDRESS,'Pennslyvania District Street No6 ABD' AS B_ADDRESS FROM DUAL UNION ALL
                    SELECT 'Onion Dist.  No 63 Kartal' AS H_ADDRESS,'Onion District No 61 Kartal' AS B_ADDRESS FROM DUAL
                )
                )

---

感谢您的帮助。我编写了如下代码并且它正在工作。

 SELECT B_ADDRESS,H_ADDRESS,B_ADDRESS_C,H_ADDRESS_C,
 UTL_MATCH.jaro_winkler_similarity(B_ADDRESS_C,H_ADDRESS_C) AS JWS,
 UTL_MATCH.jaro_winkler(B_ADDRESS_C,H_ADDRESS_C) AS JW,
 UTL_MATCH.edit_distance_similarity(B_ADDRESS_C,H_ADDRESS_C) AS EDS,
 UTL_MATCH.edit_distance(B_ADDRESS_C,H_ADDRESS_C) AS ED

 FROM (

 SELECT H_ADDRESS,B_ADDRESS,
 INITCAP(REGEXP_REPLACE(REGEXP_REPLACE(REGEXP_REPLACE(H_ADDRESS,'S[a- 
 zA-Z]{1,}|S[a-zA-Z]r|S[t]','Street'),'D[a-zA-Z]{1,}|D[a-zA-Z] 
 {1,}|D[a-zA-Z]','District'),'[.: ]',' ')) AS H_ADDRESS_C,
 INITCAP(REGEXP_REPLACE(REGEXP_REPLACE(REGEXP_REPLACE(B_ADDRESS,'S[a- 
 zA-Z]{1,}|S[a-zA-Z]r|S[t]','Street'),'D[a-zA-Z]{1,}|D[a-zA-Z] 
 {1,}|D[a-zA-Z]','District'),'[.: ]',' '))  AS B_ADDRESS_C


            FROM
            ( 
                SELECT 'Washington Str. No:60 ABD' AS 
 H_ADDRESS,'Washington Street No60 ABD' AS B_ADDRESS FROM DUAL UNION 
 ALL 
                SELECT 'Pennslyvania Dt. St. No 6 ABD' AS 
 H_ADDRESS,'Pennslyvania District Street No6 ABD' AS B_ADDRESS FROM 
 DUAL UNION ALL
                SELECT 'Onion Dist. No 63 Kartal' AS H_ADDRESS,'Onion 
 District No 61 Kartal' AS B_ADDRESS FROM DUAL
            )
            )

Answer 1

基本上，您正在寻找相似地址。有多相似，由您决定。

有样本数据：

SQL> select * from test;

H_ADDRESS                     B_ADDRESS
----------------------------- ------------------------------------
Washington Str. No:60 ABD     Washington Street No60 ABD
Pennsylvania Dt. St. No 6 ABD Pennslyvania District Street No6 ABD
Onion Dist.  No 63 Kartal     Onion District No 61 Kartal

您可以使用utl_match.jaro_winkler_similarity，它表示两个值的相似程度（百分比）。它返回

SQL> select h_address, b_address,
  2    utl_match.jaro_winkler_similarity(h_address, b_address) sim
  3  from test;

H_ADDRESS                     B_ADDRESS                                   SIM
----------------------------- ------------------------------------ ----------
Washington Str. No:60 ABD     Washington Street No60 ABD                   96
Pennsylvania Dt. St. No 6 ABD Pennslyvania District Street No6 ABD         88
Onion Dist.  No 63 Kartal     Onion District No 61 Kartal                  89

SQL>

现在，正如我所说，由您决定哪种相似性对您来说足够好。是80%吗？ 90%？ 95%？我不能说，你决定。当你这样做时，将其用作

SQL> with temp as
  2    (select h_address, b_address,
  3       utl_match.jaro_winkler_similarity(h_address, b_address) sim
  4     from test
  5    )
  6  select h_address, b_address
  7  from temp
  8  where sim >= 90;           --> here

H_ADDRESS                     B_ADDRESS
----------------------------- ------------------------------------
Washington Str. No:60 ABD     Washington Street No60 ABD

SQL>

或者，将其全部（地址和相似性）作为结果，并让其他人进行比较和过滤。