MySQL full text search equivalents to Google search: Difference between revisions

From LemonWiki共筆
Jump to navigation Jump to search
Line 53: Line 53:


== Highlight search query keywords on resulting pages ==
== Highlight search query keywords on resulting pages ==
Returned result: Show 10 characters before or after the search keywords. (Total 130 ~ 240 characters on Google resulting pages.)
Returned result: Show 10 characters before or after the search keywords. (cf: Total 130 ~ 240 characters on Google resulting pages.)


=== MySQL approach ===
=== MySQL approach ===

Revision as of 11:06, 19 October 2017

AND

Google search: keyword1 keyword2 same as keyword1 AND keyword2 or keyword1 +keyword2. Icon_exclaim.gif (1) The following is exact words search. (2) Replace column_name with your column name

  • Google: 易筋經 AND 吸星大法
  • MySQL: column_name REGEXP '易筋經' AND column_name REGEXP '吸星大法'
  • MySQL: column_name LIKE '%易筋經%' AND column_name LIKE '%吸星大法%' (online demo[1])
  • MySQL: IF(LOCATE('易筋經', column_name) > 0) AND IF(LOCATE('吸星大法, column_name) > 0)
  • MySQL: column_name LIKE '%易筋經%吸星大法%' AND column_name LIKE '%吸星大法%易筋經%' Icon_exclaim.gif Trivial for multiple keywords

OR

Google search: keyword1 OR keyword2

  • Google: 易筋經 OR 吸星大法
  • MySQL: column_name REGEXP '易筋經' OR column_name REGEXP '吸星大法'
  • MySQL: IF(LOCATE('易筋經', column_name) > 0) OR IF(LOCATE('吸星大法, column_name) > 0)
  • MySQL: column_name LIKE '%易筋經%' OR column_name LIKE '%吸星大法%' (online demo[2])

NOT

Google search: keyword1 NOT keyword2 same as keyword1 -keyword2

  • Google: 易筋經 NOT 吸星大法
  • MySQL: column_name REGEXP '易筋經' AND column_name NOT REGEXP '吸星大法' (online demo)
  • MySQL: IF(LOCATE('易筋經', column_name) > 0) AND IF(LOCATE('吸星大法, column_name) = 0)
  • MySQL: column_name LIKE '%易筋經%' AND column_name NOT LIKE '%吸星大法%'

* wildcard operator

Google * wildcard operator. "Use *, an asterisk character, known as a wildcard, to match one or more words in a phrase" [1] (online demo)

  • Google: 狐狸*叫
  • MySQL: column_name LIKE '狐狸%叫'[2]

English issue

When the keyword is short and written in English e.g. AI, the query result using column_name LIKE '%AI%' may NOT what you want e.g. Tainan, main, hair and so on.

  • (1) Remove all non-alpha-numeric-characters[3] (2) REGEXP word boundaries[4] e.g. (REPLACE(CONVERT(column_name USING ascii), '?', ' ') REGEXP '([[:<:]])AI([[:>:]])')


Cited from MySQL :: MySQL 5.7 Reference Manual :: 12.5.2 Regular Expressions

[[:<:]], [[:>:]]

These markers stand for word boundaries. They match the beginning and end of words, respectively. A word is a sequence of word characters that is not preceded by or followed by word characters. A word character is an alphanumeric character in the alnum class or an underscore (_).

Ignore special characters

Ignore return symbol and span tag

  • Example:
    • Searched the keywords e.g. "意法" site:ptt.cc on Google and found the search result contains 意 & 法 located in the nearest but different rows. 意 is at the end of the n-th row. 法 is at the beginning of n+1-th row [3].
  • Approach: (1) remove the html tag (2) remove the return symbol (Carriage return).

Ignore space, Halfwidth and fullwidth symbol (半形字元和全形字元)

  • Examples:
    • Searched the keywords e.g. "嗎有" on Google and found the search result contains 嗎? 有 & 嗎- 有.
    • Searched the keywords e.g. "人物誌Persona" on Google and found the search result contains 人物誌(Persona), 人物誌(Persona) & 「人物誌」(persona).
  • Approach: (1) remove the space symbol (2) remove the Halfwidth and fullwidth symbol.
  • References: PHP remove symbols from string - Stack Overflow

Highlight search query keywords on resulting pages

Returned result: Show 10 characters before or after the search keywords. (cf: Total 130 ~ 240 characters on Google resulting pages.)

MySQL approach

Input search keywords, and returned the matched paragraph. Using MySQL SUBSTRING() function, POSITION() function & CHAR_LENGTH() function.


SET @term := "吸星大法";
SET @message := "笑傲江湖中嵩山派掌門左冷禪所創掌法,可發出至陰至寒的真氣。左冷禪與任我行比武時,以此功對付吸星大法,使其全身凍僵、天池穴被封;與岳不群比劍奪帥時,左又使出寒冰神掌,與紫霞神功旗鼓相當、不分勝敗。

原文網址:https://kknews.cc/zh-tw/culture/xzaxbq.html";


SELECT 
@message

, CASE
  WHEN POSITION(@term IN @message) > 0 THEN SUBSTRING(@message
        , IF(
            POSITION(@term IN @message) > 0 &&
            POSITION(@term IN @message) -10 < 0
            , 1
            , POSITION(@term IN @message) -10)
        , CHAR_LENGTH(@term) + 20
      )
  ELSE ''
END AS "scrapbook"

-- Returned result of scrapbook column: Show 10 characters before or after the search keywords.
-- 行比武時,以此功對付吸星大法,使其全身凍僵、天池


SET @term := "吸星大法";
SET @message := "原文網址:https://kknews.cc/zh-tw/culture/xzaxbq.html";


SELECT 
@message

, CASE
  WHEN POSITION(@term IN @message) > 0 THEN SUBSTRING(@message
        , IF(
            POSITION(@term IN @message) > 0 &&
            POSITION(@term IN @message) -10 < 0
            , 1
            , POSITION(@term IN @message) -10)
        , CHAR_LENGTH(@term) + 20
      )
  ELSE ''
END AS "scrapbook"

-- Returned result of scrapbook column: Show 10 characters before or after the search keywords.
-- [EMPTY]

Google sheet approach

Using REGEXEXTRACT function Icon_exclaim.gif case-sensitive!:

A B
1 文章 笑傲江湖中嵩山派掌門左冷禪所創掌法,可發出至陰至寒的真氣。左冷禪與任我行比武時,以此功對付吸星大法,使其全身凍僵、天池穴被封;與岳不群比劍奪帥時,左又使出寒冰神掌,與紫霞神功旗鼓相當、不分勝敗。 原文網址:https://kknews.cc/zh-tw/culture/xzaxbq.html
2 關鍵字 吸星大法
3 搜尋結果摘要 =IF(ISERROR(REGEXEXTRACT(LOWER(B1), "(.{10}"&B2&".{10})")), "", REGEXEXTRACT(LOWER(B1), "(.{10}"&B2&".{10})"))

Microsoft Spreadsheet approach

Using FIND, MID & CONCATENATE functions

A B
1 文章 笑傲江湖中嵩山派掌門左冷禪所創掌法,可發出至陰至寒的真氣。左冷禪與任我行比武時,以此功對付吸星大法,使其全身凍僵、天池穴被封;與岳不群比劍奪帥時,左又使出寒冰神掌,與紫霞神功旗鼓相當、不分勝敗。 原文網址:https://kknews.cc/zh-tw/culture/xzaxbq.html
2 關鍵字 吸星大法
3 搜尋結果摘要 =IF(ISERROR(FIND(B2, B1)), "", CONCATENATE(MID(B1, IF(FIND(B2, B1)-10 >= 1, FIND(B2, B1)-10, 1), 10), MID(B1, FIND(B2, B1), 10+LEN(B2))))

PHP approach

PHP solution: php - highlight multiple keywords in search - Stack Overflow Unverified

Ranking factors

Possibile factors

References or related articles

to explore strange new worlds / related articles:

other search cases: if the column ... (inspired by OutWit)

  • contains ____
  • does not contain ____
  • begins with ____
  • does not begins with ____
  • ends with ____
  • does not ends with ____
  • equals to ____
  • does not equal ____

references


related news:

MySQL OR nosql related news
Failed to load RSS feed from https://news.google.com.tw/news/feeds?pz=1&cf=all&ned=us&hl=en-US&q=mysql%20OR%20nosql&cf=all&scoring=n&output=rss: Error parsing XML for RSS

Powered by Google News