Differences

This shows you the differences between two versions of the page.

--- sql_analysis [2012/01/28 02:55] – daniel
+++ sql_analysis [2016/07/21 04:00] – [Presence and count of specific mail headers] daniel
@@ Line 8: / Line 8: @@
 FROM header, (values ('X-Priority'), ('Importance'), ('Precedence'), ('Priority'), ('X-MSMail-Priority'), ('X-MS-Priority')) as h(field)
 WHERE position(E'\n'||field IN lines)>0
-group by split_part(substr(lines, 1+position(E'\n'||field in lines), 200), E'\n', 1)
+group by 1
 </code>
@@ Line 48: / Line 48: @@
  X-Priority: Normal        |     5
 </code>
+This alternative implementation uses regular expressions and differs in that it doesn't limit header values to 200 characters or any other fixed length.
+<code sql>
+select FIELD||':'||arr[1], count(*)
+ from (select FIELD, regexp_matches(lines, '(?:^|\n)' || FIELD || ':\s*([^\n]*)\n', 'g') as arr
+       from header,
+         (VALUES ('X-Priority'), ('Importance'), ('Precedence'), ('Priority'),
+             ('X-MSMail-Priority'), ('X-MS-Priority')) AS h(FIELD)
+             where strpos(lines,FIELD)>0) l
+GROUP BY 1 ORDER BY 1
+</code>
+In this version, the '' strpos(lines,FIELD)>0 '' condition is not essential: it's introduced only as a first-pass filter to eliminate the headers that don't contain anywhere any of the searched fields.
 ====== Duplicate messages ======
 This query finds each message that share the exact same headers than another message with a lower mail_id, which means that it's a duplicate.