Translate greek characters in PostgreSQL full-text search - postgresql

I would like to translate greek characters to their common latin equivalents for the purpose of full-text search.
Consider the following:
SELECT
to_tsvector('english', 'α-decay') ## to_tsquery('α & decay') AS greek_greek,
to_tsvector('english', 'α-decay') ## to_tsquery('a & decay') AS greek_latin_short,
to_tsvector('english', 'α-decay') ## to_tsquery('alpha & decay') AS greek_latin_long;
greek_greek | greek_latin_short | greek_latin_long
-------------+-------------------+------------------
t | t | f
(1 row)
The long version does not match, but users expecting these symbols might type in alpha or beta instead of α and β. Is there a pre-defined dictionary which would automatically turn α into both 'a' and 'alpha'? If not, how can I make one? Or is there a better way altogether?

You'd have to use a synonym dictionary with a synonym file like:
α alpha
β beta
ɣ gamma
...

Related

Need to explain the kdb/q script to save partitioned table

I'm trying to understand this snippet code from:
https://code.kx.com/q/kb/loading-from-large-files/
to customize it by myself (e.x partition by hours, minutes, number of ticks,...):
$ cat fs.q
\d .Q
/ extension of .Q.dpft to separate table name & data
/ and allow append or overwrite
/ pass table data in t, table name in n, : or , in g
k)dpfgnt:{[d;p;f;g;n;t]if[~&/qm'r:+en[d]t;'`unmappable];
{[d;g;t;i;x]#[d;x;g;t[x]i]}[d:par[d;p;n];g;r;<r f]'!r;
#[;f;`p#]#[d;`.d;:;f,r#&~f=r:!r];n}
/ generalization of .Q.dpfnt to auto-partition and save a multi-partition table
/ pass table data in t, table name in n, name of column to partition on in c
k)dcfgnt:{[d;c;f;g;n;t]*p dpfgnt[d;;f;g;n]'?[t;;0b;()]',:'(=;c;)'p:?[;();();c]?[t;();1b;(,c)!,c]}
\d .
r:flip`date`open`high`low`close`volume`sym!("DFFFFIS";",")0:
w:.Q.dcfgnt[`:db;`date;`sym;,;`stats]
.Q.fs[w r#]`:file.csv
But I couldn't find any resources to give me detail explain. For example:
if[~&/qm'r:+en[d]t;'`unmappable];
what does it do with the parameter d?
(Promoting this to an answer as I believe it helps answer the question).
Following on from the comment chain: in order to translate the k code into q code (or simply to understand the k code) you have a few options, none of which are particularly well documented as it defeats the purpose of the q language - to be the wrapper which obscures the k language.
Option 1 is to inspect the built-in functions in the .q namespace
q).q
| ::
neg | -:
not | ~:
null | ^:
string | $:
reciprocal| %:
floor | _:
...
Option 2 is to inspect the q.k script which creates the above namespace (be careful not to edit/change this):
vi $QHOME/q.k
Option 3 is to lookup some of the nuggets of documentation on the code.kx website, for example https://code.kx.com/q/wp/parse-trees/#k4-q-and-qk and https://code.kx.com/q/basics/exposed-infrastructure/#unary-forms
Options 4 is to google search for reference material for other/similar versions of k, for example k2/k3. They tend to be similar-ish.
Final point to note is that in most of these example you'll see a colon (:) after the primitives....this colon is required in q/kdb to use the monadic form of the primitive (most are heavily overloaded) while in k it is not required to explicitly force the monadic form. This is why where will show as &: in the q reference but will usually just be & in actual k code

Postgres Full-Text Search with Hyphen and Numerals

I have observed what seems to me an odd behavior of Postgres' to_tsvector function.
SELECT to_tsvector('english', 'abc-xyz');
returns
'abc':2 'abc-xyz':1 'xyz':3
However,
SELECT to_tsvector('english', 'abc-001');
returns
'-001':2 'abc':1
Why not something like this?
'abc':2 'abc-001':1 '001':3
And what should I do to be able to search by the numeric portion alone, without the hyphen?
Seems the text search parser identifies the hyphen followed by digits to be the sign of a signed integer. Debug with ts_debug():
SELECT * FROM ts_debug('english', 'abc-001');
alias | description | token | dictionaries | dictionary | lexemes
-----------+-----------------+-------+--------------+------------+---------
asciiword | Word, all ASCII | abc | {simple} | simple | {abc}
int | Signed integer | -001 | {simple} | simple | {-001}
Other text search configurations (like 'simple' instead of 'english') won't help as the parser itself is "at fault" here (debatable).
A simple way around it (other than modifying the parser, which I never tried) would to pre-process strings and replace hyphens with m-dash (—) or just blanks to make sure those are identified as "Space symbols". (Actual signed integers lose their negative sign in the process.)
SELECT to_tsvector('english', translate('abc-001', '-', '—'))
## to_tsquery ('english', '001'); -- true now
db<>fiddle here
This can be circumvented with PG13's dict-int addon's absval option. See the official documentation.
But in case you're stuck with an earlier PG version, here's the generalized version of a "number or negative number" workaround in a query.
select regexp_replace($$'test' & '1':* & '2'$$::tsquery::text,
'''([.\d]+''(:\*)?)', '(''\1 | ''-\1)', 'g')::tsquery;
This results in:
'test' & ( '1':* | '-1':* ) & ( '2' | '-2' )
It replaces lexemes that look like positive numbers with "number or negative number" kind of subqueries.
The double cast ::tsquery::text is just there to show how you would pass a tsquery casted to text.
Note that it handles prefix matching numeric lexemes as well.

PostgreSQL tuple format

Is there any document describing the tuple format that PostgreSQL server adheres to? The official documentation appears arcane about this.
A single tuple seems simple enough to figure out, but when it comes to arrays of tuples, arrays of composite tuples, and finally nested arrays of composite tuples, it is impossible to be certain about the format simply by looking at the output.
I am asking this following my initial attempt at implementing pg-tuple, a parser that's still missing today, to be able to parse PostgreSQL tuples within Node.js
Examples
create type type_A as (
a int,
b text
);
with a simple text: (1,hello)
with a complex text: (1,"hello world!")
create type type_B as (
c type_A,
d type_A[]
);
simple-value array: {"(2,two)","(3,three)"}
for type_B[] we can get:
{"(\"(7,inner)\",\"{\"\"(88,eight-1)\"\",\"\"(99,nine-2)\"\"}\")","(\"(77,inner)\",\"{\"\"(888,eight-3)\"\",\"\"(999,nine-4)\"\"}\")"}
It gets even more complex for multi-dimensional arrays of composite types.
UPDATE
Since it feels like there is no specification at all, I have started working on reversing it. Not sure if it can be done fully though, because from some initial examples it is often unclear what formatting rules are applied.
As Nick posted, according to docs:
the whitespace will be ignored if the field type is integer, but not
if it is text.
and
The composite output routine will put double quotes around field
values if they are empty strings or contain parentheses, commas,
double quotes, backslashes, or white space.
and
Double quotes and backslashes embedded in field values will be
doubled.
and now quoting Nick himself:
nested elements are converted to strings, and then quoted / escaped
like any other string
I give shorted example below, comfortably compared against its nested value:
a=# create table playground (t text, ta text[],f float,fa float[]);
CREATE TABLE
a=# insert into playground select 'space here',array['','bs\'],8.0,array[null,8.1];
INSERT 0 1
a=# insert into playground select 'no_space',array[null,'nospace'],9.0,array[9.1,8.0];
INSERT 0 1
a=# select playground,* from playground;
playground | t | ta | f | fa
---------------------------------------------------+------------+----------------+---+------------
("space here","{"""",""bs\\\\""}",8,"{NULL,8.1}") | space here | {"","bs\\"} | 8 | {NULL,8.1}
(no_space,"{NULL,nospace}",9,"{9.1,8}") | no_space | {NULL,nospace} | 9 | {9.1,8}
(2 rows)
If you go for deeper nested quoting, look at:
a=# select nested,* from (select playground,* from playground) nested;
nested | playground | t | ta | f | fa
-------------------------------------------------------------------------------------------------------------------------+---------------------------------------------------+------------+----------------+---+------------
("(""space here"",""{"""""""",""""bs\\\\\\\\""""}"",8,""{NULL,8.1}"")","space here","{"""",""bs\\\\""}",8,"{NULL,8.1}") | ("space here","{"""",""bs\\\\""}",8,"{NULL,8.1}") | space here | {"","bs\\"} | 8 | {NULL,8.1}
("(no_space,""{NULL,nospace}"",9,""{9.1,8}"")",no_space,"{NULL,nospace}",9,"{9.1,8}") | (no_space,"{NULL,nospace}",9,"{9.1,8}") | no_space | {NULL,nospace} | 9 | {9.1,8}
(2 rows)
As you can see, the output again follows rules the above.
This way in short answers to your questions would be:
why array is normally presented inside double-quotes, while an empty array is suddenly an open value? (text representation of empty array does not contain comma or space or etc)
why a single " is suddenly presented as \""? (text representation of 'one\ two', according to rules above is "one\\ two", and text representation of the last is ""one\\\\two"" and it is just what you get)
why unicode-formatted text is changing the escaping for \? How can we tell the difference then? (According to docs,
PostgreSQL also accepts "escape" string constants, which are an
extension to the SQL standard. An escape string constant is specified
by writing the letter E (upper or lower case) just before the opening
single quote
), so it is not unicode text, but the the way you tell postgres that it should interpret escapes in text not as symbols, but as escapes. Eg E'\'' will be interpreted as ' and '\'' will make it wait for closing ' to be interpreted. In you example E'\\ text' the text represent of it will be "\\ text" - we add backslsh for backslash and take value in double quotes - all as described in online docs.
the way that { and } are escaped is not always clear (I could not anwer this question, because it was not clear itself)

PostgreSQL full text search abbreviations

I created a Postgresql full text search using 'german'. How can I configer, that when I search for "Bezirk", lines containing "Bez." are also a match? (And vice-versa)
#pozs is right. You need to use a synonym dictionary.
1 - In the directory $SHAREDIR/tsearch_data create the file german.syn with the following contents:
Bez Bezirk
2 - Execute the query:
CREATE TEXT SEARCH DICTIONARY german_syn (
template = synonym,
synonyms = german);
CREATE TEXT SEARCH CONFIGURATION german_syn(COPY='simple');
ALTER TEXT SEARCH CONFIGURATION german_syn
ALTER MAPPING FOR asciiword, asciihword, hword_asciipart,
word, hword, hword_part
WITH german_syn, german_stem;
Now you can test it. Execute queries:
test=# SELECT to_tsvector('german_syn', 'Bezirk') ## to_tsquery('german_syn', 'Bezirk & Bez');
?column?
----------
t
(1 row)
test=# SELECT to_tsvector('german_syn', 'Bez Bez.') ## to_tsquery('german_syn', 'Bezirk');
?column?
----------
t
(1 row)
Additional links:
PostgreSQL: A Full Text Search engine (expired)
Try using a wildcard in your search.
For example:
tableName.column LIKE 'Bez%'
The % will search for any letter or number after the Bez
Description is very vague to understand what you are trying to achieve, but it looks like you need simple pattern matching search as you looking for abbreviations (so need to do stemming like in Full Text Search). I would with pg_trgm for this purpose:
WITH t(word) AS ( VALUES
('Bez'),
('Bezi'),
('Bezir')
)
SELECT word, similarity(word, 'Bezirk') AS similarity
FROM t
WHERE word % 'Bezirk'
ORDER BY similarity DESC;
Result:
word | similarity
-------+------------
Bezir | 0.625
Bezi | 0.5
Bez | 0.375
(3 rows)

Search text between symbol

I have this text (taken from concatenated field row)
Astronomic Event 2013/1434H - Aceh ....
How do We search it by 2013 or 1434h keywords?
I have tried below code but it resulting no row.
to_tsvector result:
'2013/1434h':8,12 'aceh':1 'bin.....
Sample Case:
WITH sample_table as
(SELECT to_tsvector('Astronomic Event 2013/1434H - Aceh') sample_content)
SELECT *
FROM sample_table, to_tsquery('2013') q
WHERE sample_content ## q
How do We search it by 2013 or 1434h keywords?
It seems like you want to replace:
to_tsquery('1434h') q
with:
to_tsquery('1434h | 2013') q
http://www.postgresql.org/docs/current/static/functions-textsearch.html
Side note: the to_tsquery() syntax is extremely capricious. It doesn't allow for much if any fantasy, and many of the assumptions in Postgres are everything but end-user friendly.
More often than not, you'll be better off using plainto_tsquery(), which allows any amount of garbage to be thrown at it. Thus, consider pre-processing the string before issuing the query. For instance, you could split the string, and OR the original parts together:
where sc.text_index ## (plainto_tsquery('1434h') || plainto_tsquery('2013'))
Doing so will make your code a bit more complex, but it won't rely on your users needing to understand that (contrary to what they're accustomed to in Google) they should enter 'quick & brown & fox & jumps & lazy & dog' instead of plain 'The quick brown fox jumps over the lazy dog'.
Edit: I ended up actually trying your sample query, and it seems you're actually running into a parser issue:
# SELECT alias, description, token FROM ts_debug('Astronomic Event 2013/1434H - Aceh');
alias | description | token
-----------+-------------------+------------
asciiword | Word, all ASCII | Astronomic
blank | Space symbols |
asciiword | Word, all ASCII | Event
blank | Space symbols |
file | File or path name | 2013/1434H
blank | Space symbols |
blank | Space symbols | -
asciiword | Word, all ASCII | Aceh
(8 rows)
http://www.postgresql.org/docs/current/static/textsearch-parsers.html
It looks like you might need to write (or find) and configure an app-specific parser. Having never done this personally, the best I can do is to highlight that Postgres allows this and includes a sample:
http://www.postgresql.org/docs/current/static/test-parser.html
Alternatively, change your tsvector-related trigger so that it matches e.g. \d{4}/\d+[a-zA-Z] or whatever seems most appropriate, and adds spaces accordingly, before converting it to a tsvector. Something as simple as the following might do the trick if you never need to store file names:
SELECT alias, description, token
FROM ts_debug(replace('Astronomic Event 2013/1434H - Aceh', '/', ' / '));