我有一个带有职位名称的表 例如
我正在尝试获取一些独特的短语 例如 -2 x需要 -2 x需要咖啡师 -周末2次 -2 x -每小时2 x 24 $
我已经运行了一个表格，将我的文本转换成单词数组
CREATE TABLE IF NOT EXISTS job_words (
source VARCHAR,
title VARCHAR,
words VARCHAR[]
)
我已经分割标题并作为单词插入到该表中
insert into job_words
select 'job-title', raw_title, string_to_array(raw_title, ' ') from jobs
最长的句子中有49个单词 我想找到2到10个字长的词组 如果可能的话，很乐意使用另一张表写入或直接查询
样品查询以获取一些样品数据
select cardinality(words) no_of_words, words, title from job_words
where cardinality(words) > 4 and cardinality(words) < 10 and title ilike 'need a%'
order by title limit 100
样本数据
8;"{Need,a,baby,sitter,for,an,amazing,girl}"
7;"{Need,a,baby,sitter,for,casual,sitting}"
8;"{Need,a,babysitter,for,our,19,months,old}"
9;"{Need,a,babysitter,for,our,4,year,old,son}"
9;"{Need,a,babysitter,for,our,little,reyon,19,months}"
7;"{Need,a,babysitter,-,look,no,further}"
5;"{Need,a,babysitter,or,tutor?}"
9;"{Need,a,baby,sitter,tonight,kids,are,already,sleeping}"
6;"{Need,a,Baker,now?,I'm,available!}"
6;"{Need,a,barista,all,rounder,ASAP}"
8;"{NEED,A,BARISTA???,full,time,or,part,time.}"
5;"{Need,a,brick,labourer,urgently}"
9;"{Need,a,care,giver,for,a,Month,old,baby}"
7;"{Need,a,Carer,-,After,School,hours}"
7;"{Need,a,Carpenter,-,build,a,cubby}"
5;"{Need,a,Carwash,staff,asap}"
5;"{Need,a,catering,assistant,job}"
9;"{Need,a,change,from,customer,service?,Look,no,further!}"
5;"{Need,a,change,of,scenery?}"
6;"{NEED,A,CLEANER,-,asap,start}"
6;"{Need,a,cleaner,for,daily,work}"
6;"{Need,a,cleaner,for,daily,work}"
9;"{Need,a,Cleaner,for,hotel,in,Belmont,near,Geelong}"
9;"{Need,a,Cleaner,for,hotel,in,Fyansford,near,Geelong}"
9;"{Need,a,Cleaner,for,hotel,in,Queenscliff,near,Geelong}"
5;"{Need,a,cleaner,for,mcdownal}"
7;"{Need,a,cleaner,for,tomorrow,pay,cash}"
7;"{Need,a,cleaner,for,tomorrow,pay,cash}"
5;"{Need,a,cleaner,in,Brisbane}"
6;"{Need,a,cleaner,in,Roxburghpark,Area}"
7;"{Need,a,cleaner,on,a,weekly,basis}"
7;"{Need,a,cleaner,on,Sunday,18th,June}"
9;"{Need,a,cleaning,team,for,your,building,or,office?}"
8;"{Need,a,concreter,to,start,full,time,/paving}"
6;"{Need,a,contract,climber,on,Tuesday}"
7;"{Need,a,cook,for,Road,Trip,Film}"
6;"{Need,a,delivery,driver,in,kew}"
8;"{Need,a,dishwasher,-,Wetherill,Park,6,days}"
7;"{Need,admin,done,for,hair,salon,asap}"
7;"{Need,admin,done,for,hair,salon,asap}"
7;"{Need,admin,done,for,hair,salon,asap}"
6;"{Need,a,driver,at,8:00,tonight}"
8;"{Need,a,driver,for,my,4.5,tomne,truck}"
7;"{Need,a,driver,in,a,Korean,restaurant}"
6;"{NEED,A,EXPERIENCED,CAR,WASH,STUFF}"
8;"{Need,a,flexible,babysitter,to,suit,shift,work}"
7;"{Need,a,fridge,picked,up,tommorow,Saturday}"
7;"{Need,after,school,care,with,pick,up}"
5;"{need,a,fulltime,female,cleaner}"
5;"{Need,a,full,time,job}"
8;"{Need,a,full,time,nanny,at,Baulkham,Hills}"
6;"{Need,a,"fun,","reliable,",interactive,babysitter}"
9;"{Need,a,gardener,/,labourer,tomorrow,for,5,hours}"
6;"{Need,a,gardener,or,labourer,tomorrow}"
6;"{Need,a,girl,for,sharing,room}"
8;"{Need,a,girl,or,boy,for,cleaning,job}"
6;"{Need,a,good,Barista,in,putney}"
8;"{Need,a,good,painter,for,the,next,month}"
7;"{Need,a,gyprock,setter,for,monday,23/01/17}"
9;"{Need,a,handy,person,in,our,new,work,shop}"
8;"{Need,a,helper,for,a,house,removals,truck}"
5;"{Need,a,house,Cleaner,?}"
6;"{Need,a,house,cleaner?,Call,now}"
6;"{Need,a,house,cleaner?,CALL,NOW!}"
8;"{Need,a,house,cleaner,for,this,afternoon,$30p/h}"
5;"{Need,a,housekeeper,tomorrow,morning}"
9;"{Need,a,HR,driver,for,one,day,a,week}"
8;"{need,a,invester,for,the,new,restaurant,Urgently}"
7;"{Need,a,job,asap,will,start,tomorrow}"
9;"{Need,a,job,?,Backpackers,wanted,+,FREE,ACCOMODATION}"
5;"{Need,a,job,for,weekend.}"
8;"{need,a,job,in,day,time,and,weekends}"
7;"{Need,a,job,of,cleaning,or,handkitchen}"
5;"{NEED,A,JOB?!,Start,immediately!!}"
5;"{need,a,job,(student,here)}"
6;"{Need,a,job,to,start,asap}"
9;"{Need,a,kitchen,hand,for,Indian,take,away,shop}"
8;"{Need,a,labourer.,Easy,work.,Monday,or,Tues}"
6;"{need,a,labourer,for,7,weeks}"
5;"{Need,a,labourer,for,today}"
5;"{Need,a,labourer,next,week}"
9;"{Need,a,last,minute,barista,or,chef?!,Staff,cancelled?!}"
9;"{Need,a,live,in,nanny,for,our,2,sons}"
8;"{Need,a,local,Electrician?,Look,no,further,:)}"
8;"{Need,a,male,cleaner,for,a,busy,restaurant}"
6;"{Need,a,man,and,a,ute!!!}"
9;"{NEED,A,MAN,AND,UTE,MONDAY,26th,AFTER,5PM}"
9;"{NEED,A,MANSPOWER,TO,HELP,US,IN,OUR,MOVING}"
8;"{Need,an,after-school,nanny,for,month,of,October}"
5;"{Need,a,Nanny,/,Babysitter}"
6;"{Need,a,nanny,for,2,kids}"
9;"{Need,a,nanny,for,3,days,after,school,care}"
7;"{Need,a,Nanny,for,7,year,old}"
9;"{Need,a,nanny,for,a,few,days,a,week}"
9;"{Need,a,nanny,for,after,school,pick,and,care}"
8;"{Need,a,nanny,for,immediate,start,on,Thursdays}"
9;"{Need,a,nanny,for,my,3,year,old,daughter}"
9;"{Need,a,nanny,for,my,3,year,old,daughter}"
6;"{Need,a,nanny,for,one,day.}"
7;"{Need,a,nanny,for,"Rydges,",Campbelltown,5.45-10.30pm}"
答案 0 :(得分：1)
我明白了。这是您期望的结果吗？
snprintf如果这是您的预期结果，请查询它。但是我不确定您是否应该使用庞大的数据集来执行此操作！
我从简单的短语开始，而不是从数组的示例数据开始。另外，我为每个短语添加了一个id列：
phrase_part count
------------------------------------------------------------------------- -----
{"Need","a"} 32
{"Need","a","cleaner"} 9
{"a","cleaner"} 9
{"cleaner","for"} 5
{"Need","a","babysitter"} 5
{"a","babysitter"} 5
{"a","cleaner","for"} 5
{"Need","a","cleaner","for"} 5
{"for","hotel","in"} 3
{"near","Geelong"} 3
{"hotel","in"} 3
{"Cleaner","for","hotel"} 3
{"for","hotel"} 3
{"babysitter","for","our"} 3
{"for","our"} 3
{"babysitter","for"} 3
{"a","Cleaner","for"} 3
{"Need","a","babysitter","for"} 3
...
{"NEED","A"} 2
{"a","cleaner","for","daily"} 2
{"Need","a","change"} 2
{"daily","work"} 2
{"pay","cash"} 2
{"a","cleaner","in"} 2
{"Need","a","cleaner","for","tomorrow","pay","cash"} 2
{"for","casual","sitting"} 1
{"concreter","to","start","full"} 1
{"a","change","from","customer","service?","Look"} 1
{"a","contract","climber"} 1
{"Roxburghpark","Area"} 1
{"Need","a","Carwash"} 1
...
A：将普通短语格式化为单个单词数组，并将表格扩展为每行一个单词
B：添加一个单词计数器以使用窗口函数来识别短语的第n个单词
C：将短语与自己交叉连接；更好的说：将同一个词的每个后续词连在一起
D：此窗口功能汇总短语词。它会产生类似
的结果WITH phrases as (
SELECT
*,
row_number() over (partition by id) nth_word -- B
FROM (
SELECT
id,
unnest(string_to_array(phrase, ' ')) as word -- A
FROM testdata.phrases
)s
)
SELECT
phrase_part,
count(phrase_part) FILTER (WHERE cardinality(phrase_part) >= 2) -- E
FROM (
SELECT
*,
array_agg(b.word) over (partition by a.id, a.nth_word order by a.id, a.nth_word, b.nth_word) --D
as phrase_part
FROM phrases a
JOIN phrases b -- C
ON (a.id = b.id AND a.nth_word <= b.nth_word)
) s
GROUP BY phrase_part
ORDER BY COUNT DESC
E：按短语分组并计数元素。使用filter子句可以计算不同的基数。