I'd like to make my own custom full text search and I am not sure what is the best way to make index table.
Ok, I take text field and extract all the words that are longer than 3 to the index table. But what do I have to store about them? word, ID of the table where I am searching? Anything else? Frequency of the word?
And support question: How do I split the text field to words, is there any mysql function or should I do this using server side language?
UPDATE: To make things clear: I don't need full text search just a wordlist of the words that are in all records of my text field, so I could search for the endings with LIKE 'word%'
If you are only going to implement what MySQL calls boolean mode (no relevance counting), you should implement the following basic functionality:
A wordbreaker, an algorithm that splits the strings into words. This is trivial in English but can be a problem for some Asian languages which do not use spaces between words.
Optionally, a stemmer, an algorithm which reduces words to their basic forms, so that went and gone both become go.
Optionally, a spellchecker, an algorithm which corrects the common spelling errors.
Optionally, a thesaurus, which reduces the synonyms to their common form.
A result of all this is that you have a string like this:
a fast oburn vixen jmups over an indolent canine
split into the basic forms of the words with the synonyms replaced and errors corrected:
quick
brown
fox
jump
over
lazy
dog
Then you just create a composite index on (word, rowid), where word is the basic form and rowid is the PRIMARY KEY of the record indexed.
To query for, say, '+quick +fox', you should search your index for these words and find an intersection on rowid. The intersecting rowid will contain both words.
If you are going to take relevance into account, you should additionally maintain a per-word statistics in a separate index over the whole corpus.
I should warn you that this is not a simple task. Just take a look at Sphinx source code.
Don't do it
Unless you know what you are doing forget about rolling your own full-text-search.
Let MySQL do the heavy lifting.
Use MyISAM for the table your want to search on
Put a FULLTEXT index on the text-fields you want to index.
Then do
SELECT *, MATCH(field1, field2) AGAINST 'text to search'
IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION AS relevancy
FROM table1
WHERE MATCH(field1, field2) AGAINST 'text to search'
IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION
ORDER BY relevancy
See: http://dev.mysql.com/doc/refman/5.5/en/fulltext-search.html#function_match
OP indicated that he want to search for word endings.
Where I live (Holland) we do poems in December, so I do a lot of word ending searching to find words that ryhme.
Here's my trick.
Add a new field to your table named visa_versa: varchar indexed
UPDATE mytable SET mytable.visa_versa = REVERSE(mytable.myword);
Now you can do an indexed search on word endings with
SET #ending = 'end';
SELECT myword FROM mytable where visa_versa LIKE REVERSE(CONCAT('%',#ending));
Related
However I found some threads about this, but nothing fits to my case.
I have a search field in my mobile app, where after text change, the real time search is running via calling my API.
The search request starts only if there are 3 or more characters entered and is searching ONLY in 1 DB column, called TITLE. So each time the user enters a letter, a query is searching for it.
Currently I have it like this (I know this solution is very bad). $searchedword is the word user entered:
if (!empty($searchedword)&&strlen($searchedword)>2 ) {$searchedword=strtolower($searchedword);
$sql = "SELECT * FROM TABLE ";$result = $mysqli->query($sql); $output='';
if ($result->num_rows > 0) {
while($data=$result->fetch_array()) {
$title=strtolower($data['title']);$content=$data['content'];
if (strpos($title,$searchedword) !== false ) {$output.=$title.','.$content;}
}}
So this just checks, if the title from DB contains the searched word. This works very well, but I think it is very bad according to performance, because each time the user enters a letter to the search field, each time all the data from the table are queried and looked for that word.
I want to recreate my code to meet the best performance.
So my first question is, should I add a FULLTEXT INDEX to the TITLE column in DB, will it help or will it just increase the disk space? As I am just searching against 1 column and in this column is just a title (1 or 2 words max).
And second question, what should be the best query for my case and of course with the best performance? As I need to search after each letter which user enters.
Can I use the search this way?
SELECT * FROM TABLE WHERE MATCH (title) AGAINST ('$searchedword' IN NATURAL LANGUAGE MODE)
However it seems, this will return only if the word completely matches the title, but returns nothing when the word is part of the title, so it is not a good solution.
The only solution which works is this:
SELECT * FROM TABLE WHERE title LIKE '%$searchedword%' "
but what about performance? And I don't understand how this works, because searchedword are converted to lowercase and I have removed the accents from that word, and the TITLE column in DB has accents and also Uppercase, but this search works very well!
If your title column has a collation like utfmb4_general_ci, you don't have to worry about dealing with upper case, lower case, and diacritical marks in your MySQL WHERE clauses. MySQL will do it for you. It is really good at handling character sets and collations in all kinds of languages. (Such things are very helpful to Swedish-language users, and the inventors of MySQL are Swedish.)
FULLTEXT with NATURAL LANGUAGE MODE is probably not the right approach for this application. It works on words, not chunks of letters. So it probably won't give you anything until your user has typed a whole word, and not a stop word. And, it is a little squirrely when you search a table with only a few rows. So, that might be a problem if you're just getting started.
It does order the results by the closeness of the match, so the most likely hit is the first one. So, if you know you have a phrase to search, it's good.
For your progressive-search application you may want to use one of these two LIKE queries.
SELECT title FROM tbl WHERE title LIKE CONCAT('$searchedword', '%') /*insecure*/
or this one which is much slower but finds your partial match anywhere in the title, not just at the beginning.
SELECT title FROM tbl WHERE title LIKE CONCAT('%', '$searchedword', '%') /*insecure*/
Avoid running these queries until you have gathered at least a few letters from your user, otherwise you'll get absurdly many results.
In these cases say SELECT title not SELECT *, and create an ordinary index on the title column. That way MySQL can satisfy the whole query from the index, which will make it much faster.
And, use MySQL's WHERE functionality to do the matching. Don't fetch the whole table from MySQL and search it in your php program.
And, use prepared statements. Because cybercreeps.
I have strings like the following in my VARCHAR InnoDB table column:
"This is a {{aaaa->bbb->cccc}} and that is a {{dddd}}!"
Now, I'd like to search for e.g. {{xxx->yyy->zzz}}. Brackets are part of the string. Sometimes searched together with another colum, but which only contains an ordinary id and hence don't need to be considered (I guess).
I know I can use LIKE or REGEXP. But these (already tried) ways are too slow. Can I introduce a fulltext index? Or should I add another helping table? Should I replace the special characters {, }, -, > to get words for the fulltext search? Or what else could I do?
The search works with some ten-thousand rows and I assume that I often get about one hundred hits.
This link should give you all the info you need regarding FULLTEXT indexes in MySQL.
MySQL dev site
The section that you will want to pay particular attention to is:
"Full-text searching is performed using MATCH() ... AGAINST syntax. MATCH() takes a comma-separated list that names the columns to be searched. AGAINST takes a string to search for, and an optional modifier that indicates what type of search to perform. The search string must be a string value that is constant during query evaluation. This rules out, for example, a table column because that can differ for each row."
So in short, to answer your question you should see an improvement in query execution times by implementing a full text index on wide VARCHAR columns. Providing you are using a compatible storage engine ( InnoDB or MyISAM)
Also here is an example of how you can query the full text index and also an additional ID field as hinted in your question:
SELECT *
FROM table
WHERE MATCH (fieldlist) AGAINST ('search text here')
AND ( field2= '1234');
I am building a search feature for the messages part of my site, and have a messages database with a little over 9,000,000 rows, and and index on the sender, subject, and message fields. I was hoping to use the LIKE mysql clause in my query, such as (ex)
SELECT sender, subject, message FROM Messages WHERE message LIKE '%EXAMPLE_QUERY%';
to retrieve results. unfortunately, MySQL doesn't use indexes when a leading wildcard is present , and this is necessary for the search query could appear anywhere in the message (this is how the wildcards work, no?). Queries are very very slow and I cannot use a full text index either, because of the annoying 50% rule (I just can't afford to rule that much out). Is there anyway (or even, any alternative to this) to optimize a query using like and two wildcards? Any help is appreciated.
You should either use full-text indexes (you said you can't), design a full-text search by yourself or offload the search from MySQL and use Sphinx/Lucene. For Lucene you can use Zend_Search_Lucene implementation from Zend Framework or use Solr.
Normal indexes in MySQL are B+Trees, and they can't be used if the starting of the string is not known (and this is the case when you have wildcard in the beginning)
Another option is to implement search on your own, using reference table. Split text in words and create table that contains word, record_id. Then in the search you split the query in words and search for each of the words in the reference table. In this way you are not limitting yourself to the beginning of the whole text, but only to the beginning of the given word (and you'll match the rest of the words anyway)
'%EXAMPLE_QUERY%'; is a very very bad idea .. am going to give you some
A. Avoid wildcards at the start of LIKE queries use 'EXAMPLE_QUERY%'; instead
B. Create Keywords where you can easily use MATCH
If you want to stick with using MySQL, you should use FULL TEXT indexes. Full text indexes index words in a text block. You can then search on word stems and return the results in order of relevance. So you can find the word "example" within a block of text, but you still can't search efficiently on "xampl" to find "example".
MySQL's full text search is not great, but it is functional.
http://dev.mysql.com/doc/refman/5.1/en/fulltext-search.html
select * from emp where ename like '%e';
gives emp_name that ends with letter e.
select * from emp where ename like 'A%';
gives emp_name that begins with letter a.
select * from emp where ename like '_a%';
gives emp_name in which second letter is a.
I have a column in my table called tags. It has comma delimited text in it (box, hat, car).
I select a row from my table and I want to find other rows that have similar tags to the tags in the row selected.
I know it could be a better table design, but I can't change the design.
I know this will search the tags for a keyword, but I don't want to search via a keyword but by a list of tags.
WHERE (',' + Tags + ',') LIKE '%,keyword,%'
Does anyone know how I would do this?
Using MYSQL.
You say you can't change the design. Can you add a FULLTEXT index and put this table in the MyISAM access method? If so you can use FULLTEXT searching. For your application you'll do best using BOOLEAN mode.
WHERE MATCH (tags)
AGAINST ('box hat -car' IN BOOLEAN MODE);
This particular search looks for keywords box and hat, and excludes the keyword car.
Here's a description. This might work well for you.
http://dev.mysql.com/doc/refman/5.5/en/fulltext-boolean.html
You can run a FULLTEXT search in InnoDB. But I don't believe you can index it.
If you can't use a fulltext search, you're stuck with the wrong tool for your job. Matching comma-delimited strings with SQL is like driving nails by hitting them with the handle of a screwdriver. Both take a long time and are incredibly awkward.
I'm recently doing some string searches from a table with about 50k strings in it, fairly large I'd say but not that big. I was doing some nested queries for a 'search within results' kinda thing. I was using LIKE statement to get a match of a searched keyword.
I came across MySQL's Full-Text search which I tried so I added a fulltext index to my str column. I'm aware that Full-text searches doesn't work on virtually created tables or even with Views so queries with sub-selects will not fit. I mentioned I was doing a nested queries, example is:
SELECT s2.id, s2.str
FROM
(
SELECT s1.id, s1.str
FROM
(
SELECT id, str
FROM strings
WHERE str LIKE '%term%'
) AS s1
WHERE s1.str LIKE '%another_term%'
) AS s2
WHERE s2.str LIKE '%a_much_deeper_term%';
This is actually not applied to any code yet, I was just doing some tests. Also, searching strings like this can be easily achieved by using Sphinx (performance wise) but let's consider Sphinx not being available and I want to know how this will work well in pure SQL query. Running this query on a table without Full-text added takes about 2.97 secs. (depends on the search term). However, running this query on a table with Full-text added to the str column finished in like 104ms which is fast (i think?).
My question is simple, is it valid to use LIKE or is it a good practice to use it at all in a table with Full-text added when normally we would use MATCH and AGAINST statements?
Thanks!
In this case you not neccessarily need subselects. You can siply use:
SELECT id, str
FROM item_strings
WHERE str LIKE '%term%'
AND str LIKE '%another_term%'
AND str LIKE '%a_much_deeper_term%'
... but also raises a good question: the order in which you are excluding the rows. I guess MySQL is smart enough to assume that the longest term will be the most restrictive, so starting with a_much_deeper_term it will eliminate most of the records then perform addtitional comparsion only on a few rows. - Contrary to this, if you start with term you will probably end up with many possible records then you have to compare them against the st of the terms.
The interesting part is that you can force the order in which the comparsion is made by using your original subselect example. This gives the opportunity to make a decision which term is the most restrictive based upon more han just the length, but for example:
the ratio of consonants a vowels
the longest chain of consonants of the word
the most used vowel in the word
...etc. You can also apply some heuristics based on the type of textual infomation you are handling.
Edit:
This is just a hunch but it could be possible to apply the LIKE to the words in the fulltext indexitself. Then match the rows against the index as if you have serched for full words.
I'm not sure if this is actually done, but it would be a smart thing to pull off by the MySQL people. Also note that this theory can only be used if all possible ocurrences arein fact in the fulltext search. For this you need that:
Your search pattern must be at least the size of the miimal word-length. (If you re searching for example %id% then it can be a part of a 3 letter word too, which is excluded by default form FULLTEXT index).
Your search pattern must not be a substring of any listed excluded word for example: and, of etc.
Your pattern must not contain any special characters.