Executable 和 ExecutablePool 表引擎

Executable(script_name, format, [input_query...])

#!/usr/bin/python3

import sys
import string
import random

def main():

    # 读取输入值
    for number in sys.stdin:
        i = int(number)

        # 生成一些随机行
        for id in range(0, i):
            letters = string.ascii_letters
            random_string =  ''.join(random.choices(letters ,k=10))
            print(str(id) + '\t' + random_string + '\n', end='')

        # 将结果输出到标准输出
        sys.stdout.flush()

if __name__ == "__main__":
    main()

CREATE TABLE my_executable_table (
   x UInt32,
   y String
)
ENGINE = Executable('my_script.py', TabSeparated, (SELECT 10))

SELECT * FROM my_executable_table

┌─x─┬─y──────────┐
│ 0 │ BsnKBsNGNH │
│ 1 │ mgHfBCUrWM │
│ 2 │ iDQAVhlygr │
│ 3 │ uNGwDuXyCk │
│ 4 │ GcFdQWvoLB │
│ 5 │ UkciuuOTVO │
│ 6 │ HoKeCdHkbs │
│ 7 │ xRvySxqAcR │
│ 8 │ LKbXPHpyDI │
│ 9 │ zxogHTzEVV │
└───┴────────────┘

CREATE TABLE sentiment (
   id UInt64,
   sentiment Float32
)
ENGINE = Executable(
    'sentiment.py',
    TabSeparated,
    (SELECT id, comment FROM hackernews WHERE id > 0 AND comment != '' LIMIT 20)
);

#!/usr/local/bin/python3.9

import sys
import nltk
from nltk.sentiment import SentimentIntensityAnalyzer

def main():
    sentiment_analyzer = SentimentIntensityAnalyzer()

    while True:
        try:
            row = sys.stdin.readline()
            if row == '':
                break

            split_line = row.split("\t")

            id = str(split_line[0])
            comment = split_line[1]

            score = sentiment_analyzer.polarity_scores(comment)['compound']
            print(id + '\t' + str(score) + '\n', end='')
            sys.stdout.flush()
        except BaseException as x:
            break

if __name__ == "__main__":
    main()

SELECT *
FROM sentiment

┌───────id─┬─sentiment─┐
│  7398199 │    0.4404 │
│ 21640317 │    0.1779 │
│ 21462000 │         0 │
│ 25168863 │         0 │
│ 25168978 │   -0.1531 │
│ 25169359 │         0 │
│ 25169394 │   -0.9231 │
│ 25169766 │    0.4137 │
│ 25172570 │    0.7469 │
│ 25173687 │    0.6249 │
│ 28291534 │         0 │
│ 28291669 │   -0.4767 │
│ 28291731 │         0 │
│ 28291949 │   -0.4767 │
│ 28292004 │    0.3612 │
│ 28292050 │    -0.296 │
│ 28292322 │         0 │
│ 28295172 │    0.7717 │
│ 28295288 │    0.4404 │
│ 21465723 │   -0.6956 │
└──────────┴───────────┘

CREATE TABLE sentiment_pooled (
   id UInt64,
   sentiment Float32
)
ENGINE = ExecutablePool(
    'sentiment.py',
    TabSeparated,
    (SELECT id, comment FROM hackernews WHERE id > 0 AND comment != '' LIMIT 20000)
)
SETTINGS
    pool_size = 4;