连接越多，PSQL 查询的结果越少

Question

我有以下 psql 表。实际上，它大约有 20 亿行。

 id  word      lemma     pos              textid  country_genre     
 1  Stuffing   stuff      vvg             190568  AN         
 2  her        her        appge           190568  AN         
 3  key        key        nn1             190568  AN         
 4  into       into       ii              190568  AN         
 5  the        the        at              190568  AN         
 6  lock       lock       nn1             190568  AN         
 7  she        she        appge           190568  AN         
 8  pushed     push       vvd             190568  AN         
 9  her        her        appge           190568  AN         
10  way        way        nn1             190568  AN         
11  into       into       ii              190568  AN         
12  the        the        appge           190568  AN         
13  house      house      nn1             190568  AN         
14  .                     .               190568  AN         
15  She        she        appge           190568  AN         
16  had        have       vhd             190568  AN         
17  also       also       rr              190568  AN         
18  cajoled    cajole     vvd             190568  AN         
19  her        her        appge           190568  AN         
20  way        way        nn1             190568  AN         
21  into       into       ii              190568  AN         
22  the        the        at              190568  AN         
23  home       home       nn1             190568  AN         
24  .                     .               190568  AN         
..  ...        ...        ..              ...     ..

我想创建下表，它并排显示了所有带有单词的“方式”成语以及“country_genre”、“lemma”和“pos”列中的一些数据。

country_genre word   word       word       lemma      pos        word       word     word       word       word       lemma      pos        word       word       
AN         lock   she        pushed     push       vvd        her        way      into       the        house      house      nn1        .          she
AN         had    also       cajoled    cajole     vvd        her        way      into       the        home       home       nn1        .          A          
AN         tried  to         force      force      vvi        her        way      into       the        palace     palace     nn1        ,          officials  

我使用以下代码（感谢 Bohemian：https://stackoverflow.com/a/47496945/3957383!）：

copy(

 SELECT
   c1.id, c1.country_genre, c1.textid, c1.wordid, c1.word,  c2.word, c3.word,  c4.word, c4.lemma, c4.pos, c5.word, c6.word, c7.word, c8.word, c9.word, c9.lemma, c9.pos, c10.word, c11.word

 FROM

 orderedflatcorpus AS c1
 JOIN orderedflatcorpus AS c2 ON c1.id + 1 = c2.id
 JOIN orderedflatcorpus AS c3 ON c1.id + 2 = c3.id 
 JOIN orderedflatcorpus AS c4 ON c1.id + 3 = c4.id
 JOIN orderedflatcorpus AS c5 ON c1.id + 4 = c5.id
 JOIN orderedflatcorpus AS c6 ON c1.id + 5 = c6.id
 JOIN orderedflatcorpus AS c7 ON c1.id + 6 = c7.id
 JOIN orderedflatcorpus AS c8 ON c1.id + 7 = c8.id
 JOIN orderedflatcorpus AS c9 ON c1.id + 8 = c9.id
 JOIN orderedflatcorpus AS c10 ON c1.id + 9 = c10.id
 JOIN orderedflatcorpus AS c11 ON c1.id + 10 = c11.id

 WHERE

 c4.pos LIKE 'vv%'
 AND c5.pos = 'appge'
 AND c6.word = 'way'
 AND c7.pos LIKE 'i%'
 AND c8.word = 'the'
 AND c9.pos LIKE 'n%'
 )

 TO

 '/home/postgres/Results/OUTPUT.csv'
 DELIMITER E'\t'
 csv header;

此查询返回 18706 个相关结构。

但是，如果我使用下面的代码，它提取了更多的上下文（21 个而不是 11 个单词），但在其他方面与前一个相同，就会发生令人担忧的事情：我只得到 18555 个相关结构。

 copy(
 SELECT c1.id, c1.country_genre, c1.textid, c1.wordid, c1.word, c1.pos, c2.word, c2.pos, c3.word, c3.pos, c4.word, c4.pos, c5.word, c5.pos, c6.word, c6.pos, 
 c7.word, c7.pos, c8.word, c8.pos, c8.lemma, c9.word, c9.pos, c10.word, c10.pos, c11.word, c11.pos, c12.word, c12.pos, c13.word, c13.pos, c13.lemma, c14.word, 
 c14.pos, c15.word, c15.pos, c16.word, c16.pos, c17.word, c17.pos, c18.word, c18.pos, c19.word, c19.pos, c20.word, c20.pos, c21.word, c21.pos 

 FROM 

 orderedflatcorpus AS c1 
 JOIN orderedflatcorpus AS c2 ON c1.id + 1 = c2.id 
 JOIN orderedflatcorpus AS c3 ON c1.id + 2 = c3.id 
 JOIN orderedflatcorpus AS c4 ON c1.id + 3 = c4.id 
 JOIN orderedflatcorpus AS c5 ON c1.id + 4 = c5.id 
 JOIN orderedflatcorpus AS c6 ON c1.id + 5 = c6.id 
 JOIN orderedflatcorpus AS c7 ON c1.id + 6 = c7.id 
 JOIN orderedflatcorpus AS c8 ON c1.id + 7 = c8.id 
 JOIN orderedflatcorpus AS c9 ON c1.id + 8 = c9.id 
 JOIN orderedflatcorpus AS c10 ON c1.id + 9 = c10.id 
 JOIN orderedflatcorpus AS c11 ON c1.id + 10 = c11.id 
 JOIN orderedflatcorpus AS c12 ON c1.id + 11 = c12.id 
 JOIN orderedflatcorpus AS c13 ON c1.id + 12 = c13.id 
 JOIN orderedflatcorpus AS c14 ON c1.id + 13 = c14.id 
 JOIN orderedflatcorpus AS c15 ON c1.id + 14 = c15.id 
 JOIN orderedflatcorpus AS c16 ON c1.id + 15 = c16.id 
 JOIN orderedflatcorpus AS c17 ON c1.id + 16 = c17.id 
 JOIN orderedflatcorpus AS c18 ON c1.id + 17 = c18.id 
 JOIN orderedflatcorpus AS c19 ON c1.id + 18 = c19.id 
 JOIN orderedflatcorpus AS c20 ON c1.id + 19 = c20.id 
 JOIN orderedflatcorpus AS c21 ON c1.id + 20 = c21.id 

 WHERE 

 c8.pos LIKE 'vv%' 
 AND c9.pos = 'appge' 
 AND c10.word = 'way' 
 AND c11.pos LIKE 'i%' 
 AND c12.word = 'the' 
 AND c13.pos LIKE 'n%' 
 ) 
 TO '/home/postgres/Results/OUTPUT.csv' DELIMITER E'\t' csv header;

我查看了第二个查询中缺少的行，但我无法检测到遗漏的任何模式。

有人知道这里会发生什么吗？谢谢！

Answer 1

在Postgres JOIN 中等价于INNER JOIN。例如，请参阅this。

换句话说，您的查询需要匹配所有ON 谓词才能得到结果。添加更多 JOINS 也会增加无法匹配的 ON 谓词，从而减少结果数量。

尝试将JOIN 更改为LEFT JOIN（又名LEFT OUTER JOIN），看看是否能获得更多结果。

如果没有看到您的示例数据和结果集中缺少的示例，很难说还有什么问题。

Answer 2

Pirho 的回答是正确的——就目前而言。但是，仅将联接更改为外部联接可能无法解决您的问题。

首先，两个查询的where条件不同，所以我没有理由假设它们会返回相同的结果集。

第二个查询是连续查找 21 个单词——不是 19 个单词，也不是 13 个单词。因此，如果模式位于文档的末尾，将不会被找到。

根据您的条件，需要连续出现 13 个单词（以达到您的条件的计数“13”。然后，我猜是可选的 7 行。这需要内部和外连接：

FROM orderedflatcorpus c1 JOIN
     orderedflatcorpus c2
     ON c1.id + 1 = c2.id JOIN
     orderedflatcorpus c3
     ON c1.id + 2 = c3.id JOIN
     orderedflatcorpus c4
     ON c1.id + 3 = c4.id JOIN
     orderedflatcorpus c5
     ON c1.id + 4 = c5.id JOIN
     orderedflatcorpus c6
     ON c1.id + 5 = c6.id JOIN
     orderedflatcorpus c7
     ON c1.id + 6 = c7.id JOIN
     orderedflatcorpus c8
     ON c1.id + 7 = c8.id JOIN
     orderedflatcorpus c9
     ON c1.id + 8 = c9.id JOIN
     orderedflatcorpus c10
     ON c1.id + 9 = c10.id JOIN
     orderedflatcorpus c11
     ON c1.id + 10 = c11.id JOIN
     orderedflatcorpus c12
     ON c1.id + 11 = c12.id JOIN
     orderedflatcorpus c13
     ON c1.id + 12 = c13.id 
     orderedflatcorpus c14
     ON c1.id + 13 = c14.id LEFT JOIN
     orderedflatcorpus c15
     ON c1.id + 14 = c15.id LEFT JOIN
     orderedflatcorpus c16
     ON c1.id + 15 = c16.id LEFT JOIN
     orderedflatcorpus c17
     ON c1.id + 16 = c17.id LEFT JOIN
     orderedflatcorpus c18
     ON c1.id + 17 = c18.id LEFT JOIN
     orderedflatcorpus c19
     ON c1.id + 18 = c19.id LEFT JOIN
     orderedflatcorpus c20
     ON c1.id + 19 = c20.id LEFT JOIN
     orderedflatcorpus c21
     ON c1.id + 20 = c21.id 
WHERE c8.pos LIKE 'vv%' AND
      c9.pos = 'appge' AND
      c10.word = 'way' AND
      c11.pos LIKE 'i%' AND
      c12.word = 'the' AND
      c13.pos LIKE 'n%' ;

下一步：我怀疑country 和/或textid 应该是join 条件的一部分。