你可以用一个 join 以及一些窗口功能：

select top (1) with ties m.masterid, paperid, version, papername,
       max(authorname) over (partition by paperid, grp) as authorname
from (select m.*, pa.authorname,
             count(pa.authorname) over (partition by m.paperid order by m.version) as grp
      from master m left join
           paperauthor pa
           on pa.masterid = m.masterid
     ) m
order by row_number() over (partition by m.paperid order by m.version desc);

它使用 with ties 获取最新版本而不需要附加子查询的技巧。
这是一把小提琴。

用下列方法尝试 row_number ，这是演示。

;with mst as
(
  select
    *,
    row_number() over (partition by PaperID order by version desc) as rnk
  from master
),
ppauth as
(
  select 
    *,
    row_number() over (order by MasterID desc) as rnk
  from PaperAuthor
)

select
  m.MasterID,
  PaperID,
  Version,
  PaperName,
  AuthorName
from mst m
join ppauth p
on m.rnk = p.rnk
where m.rnk = 1

输出：

-------------------------------------------------------------
| MasterID  PaperID  Version     PaperName        AuthorName|
-------------------------------------------------------------
|  3        ISBN-1X    3    Medical Research(2.1)    XYZ    | 
-------------------------------------------------------------

你可以试试下面的查询。它将最近的作者与每个isbn的最新图书数据对应起来。

;WITH CTE_Master AS
(
SELECT *, ROW_NUMBER() OVER(PARTITION BY PaperID ORDER BY Version DESC) AS recentOrder
FROM Master
)
SELECT c.MasterId, c.PaperId, c.Version, c.PaperName, pa.AuthorName
FROM CTE_Master as c
CROSS APPLY
(SELECT TOP 1 MasterId, AuthorName From PaperAuthor WHERE MasterId <= c.MasterId ORDER BY MasterId Desc) as pa(MasterId,AuthorName)
WHERE c.recentOrder = 1

+----------+---------+---------+-----------------------+------------+
| MasterId | PaperId | Version |       PaperName       | AuthorName |
+----------+---------+---------+-----------------------+------------+
|        3 | ISBN-1X |       3 | Medical Research(2.1) | XYZ        |
|        4 | ISBN-2X |       1 | Medical Research(2.1) | LMN        |
+----------+---------+---------+-----------------------+------------+

更新：如果你想要空作者也

;WITH CTE_Master AS
(
SELECT *, ROW_NUMBER() OVER(PARTITION BY PaperID ORDER BY Version DESC) AS recentOrder
FROM Master
)
SELECT c.MasterId, c.PaperId, c.Version, c.PaperName, pa.AuthorName
FROM CTE_Master as c
CROSS APPLY
(SELECT TOP 1 MasterId, AuthorName From PaperAuthor WHERE MasterId <= c.MasterId ORDER BY MasterId Desc) as pa(MasterId,AuthorName)
WHERE c.recentOrder = 1 and c.MasterId <= (SELECT MAX(MasterId) FROM PaperAuthor)
UNION ALL
SELECT c.MasterId, c.PaperId, c.Version, c.PaperName, NULL as AuthorName
FROM CTE_Master as c
WHERE c.MasterId > (SELECT MAX(MasterId) FROM PaperAuthor)

+---------+---------+---------+---------------------------+------------+
| asterId | PaperId | Version |         PaperName         | AuthorName |
+---------+---------+---------+---------------------------+------------+
|       3 | ISBN-1X |       3 | Medical Research(2.1)     | XYZ        |
|       4 | ISBN-2X |       1 | Medical Research(2.1)     | LMN        |
|       5 | ISBN-3X |       1 | Unpublished Research(0.0) | null       |
+---------+---------+---------+---------------------------+------------+