下载数据集
wget https://s3.amazonaws.com/menusdata.nypl.org/gzips/2021_08_01_07_01_17_data.tgz
# 可选：验证校验和
md5sum 2021_08_01_07_01_17_data.tgz
# 校验和应为：db6126724de939a5481e3160a2d67d15
解压数据集
tar xvf 2021_08_01_07_01_17_data.tgz
Menu— 菜单相关信息：餐厅名称、菜单被记录到的日期等。Dish— 菜品相关信息：菜名以及一些特征。MenuPage— 菜单页面相关信息，因为每个页面都属于某个菜单。MenuItem— 菜单中的一个条目，即某个菜单页面上的一道菜及其价格：关联到菜品和菜单页面。
创建表
CREATE TABLE dish
(
id UInt32,
name String,
description String,
menus_appeared UInt32,
times_appeared Int32,
first_appeared UInt16,
last_appeared UInt16,
lowest_price Decimal64(3),
highest_price Decimal64(3)
) ENGINE = MergeTree ORDER BY id;
CREATE TABLE menu
(
id UInt32,
name String,
sponsor String,
event String,
venue String,
place String,
physical_description String,
occasion String,
notes String,
call_number String,
keywords String,
language String,
date String,
location String,
location_type String,
currency String,
currency_symbol String,
status String,
page_count UInt16,
dish_count UInt16
) ENGINE = MergeTree ORDER BY id;
CREATE TABLE menu_page
(
id UInt32,
menu_id UInt32,
page_number UInt16,
image_id String,
full_height UInt16,
full_width UInt16,
uuid UUID
) ENGINE = MergeTree ORDER BY id;
CREATE TABLE menu_item
(
id UInt32,
menu_page_id UInt32,
price Decimal64(3),
high_price Decimal64(3),
dish_id UInt32,
created_at DateTime,
updated_at DateTime,
xpos Float64,
ypos Float64
) ENGINE = MergeTree ORDER BY id;
导入数据
clickhouse-client --format_csv_allow_single_quotes 0 --input_format_null_as_default 0 --query "INSERT INTO dish FORMAT CSVWithNames" < Dish.csv
clickhouse-client --format_csv_allow_single_quotes 0 --input_format_null_as_default 0 --query "INSERT INTO menu FORMAT CSVWithNames" < Menu.csv
clickhouse-client --format_csv_allow_single_quotes 0 --input_format_null_as_default 0 --query "INSERT INTO menu_page FORMAT CSVWithNames" < MenuPage.csv
clickhouse-client --format_csv_allow_single_quotes 0 --input_format_null_as_default 0 --date_time_input_format best_effort --query "INSERT INTO menu_item FORMAT CSVWithNames" < MenuItem.csv
format_csv_allow_single_quotes，因为数据字段只使用双引号，而单引号可以出现在值中，不应让 CSV parser 产生误判。
我们禁用 input_format_null_as_default，因为我们的数据中没有 NULL。否则 ClickHouse 会尝试解析 \N 序列，从而可能与数据中的 \ 混淆。
设置 date_time_input_format best_effort 允许以多种格式解析 DateTime 字段。例如，不带秒的 ISO-8601 格式 (如 ‘2000-01-01 01:02’) 也能被识别。如果不启用此设置，则只允许固定的 DateTime 格式。
对数据进行反规范化
JOIN。这类数据称为“反规范化的”数据。
我们将创建一个表 menu_item_denorm，其中包含通过 JOIN 合并在一起的所有数据：
CREATE TABLE menu_item_denorm
ENGINE = MergeTree ORDER BY (dish_name, created_at)
AS SELECT
price,
high_price,
created_at,
updated_at,
xpos,
ypos,
dish.id AS dish_id,
dish.name AS dish_name,
dish.description AS dish_description,
dish.menus_appeared AS dish_menus_appeared,
dish.times_appeared AS dish_times_appeared,
dish.first_appeared AS dish_first_appeared,
dish.last_appeared AS dish_last_appeared,
dish.lowest_price AS dish_lowest_price,
dish.highest_price AS dish_highest_price,
menu.id AS menu_id,
menu.name AS menu_name,
menu.sponsor AS menu_sponsor,
menu.event AS menu_event,
menu.venue AS menu_venue,
menu.place AS menu_place,
menu.physical_description AS menu_physical_description,
menu.occasion AS menu_occasion,
menu.notes AS menu_notes,
menu.call_number AS menu_call_number,
menu.keywords AS menu_keywords,
menu.language AS menu_language,
menu.date AS menu_date,
menu.location AS menu_location,
menu.location_type AS menu_location_type,
menu.currency AS menu_currency,
menu.currency_symbol AS menu_currency_symbol,
menu.status AS menu_status,
menu.page_count AS menu_page_count,
menu.dish_count AS menu_dish_count
FROM menu_item
JOIN dish ON menu_item.dish_id = dish.id
JOIN menu_page ON menu_item.menu_page_id = menu_page.id
JOIN menu ON menu_page.menu_id = menu.id;
验证数据
Query
SELECT count() FROM menu_item_denorm;
Response
┌─count()─┐
│ 1329175 │
└─────────┘
执行一些查询
菜品历史价格的平均值
Query
SELECT
round(toUInt32OrZero(extract(menu_date, '^\\d{4}')), -1) AS d,
count(),
round(avg(price), 2),
bar(avg(price), 0, 100, 100)
FROM menu_item_denorm
WHERE (menu_currency = 'Dollars') AND (d > 0) AND (d < 2022)
GROUP BY d
ORDER BY d ASC;
Response
┌────d─┬─count()─┬─round(avg(price), 2)─┬─bar(avg(price), 0, 100, 100)─┐
│ 1850 │ 618 │ 1.5 │ █▍ │
│ 1860 │ 1634 │ 1.29 │ █▎ │
│ 1870 │ 2215 │ 1.36 │ █▎ │
│ 1880 │ 3909 │ 1.01 │ █ │
│ 1890 │ 8837 │ 1.4 │ █▍ │
│ 1900 │ 176292 │ 0.68 │ ▋ │
│ 1910 │ 212196 │ 0.88 │ ▊ │
│ 1920 │ 179590 │ 0.74 │ ▋ │
│ 1930 │ 73707 │ 0.6 │ ▌ │
│ 1940 │ 58795 │ 0.57 │ ▌ │
│ 1950 │ 41407 │ 0.95 │ ▊ │
│ 1960 │ 51179 │ 1.32 │ █▎ │
│ 1970 │ 12914 │ 1.86 │ █▋ │
│ 1980 │ 7268 │ 4.35 │ ████▎ │
│ 1990 │ 11055 │ 6.03 │ ██████ │
│ 2000 │ 2467 │ 11.85 │ ███████████▋ │
│ 2010 │ 597 │ 25.66 │ █████████████████████████▋ │
└──────┴─────────┴──────────────────────┴──────────────────────────────┘
汉堡价格
Query
SELECT
round(toUInt32OrZero(extract(menu_date, '^\\d{4}')), -1) AS d,
count(),
round(avg(price), 2),
bar(avg(price), 0, 50, 100)
FROM menu_item_denorm
WHERE (menu_currency = 'Dollars') AND (d > 0) AND (d < 2022) AND (dish_name ILIKE '%burger%')
GROUP BY d
ORDER BY d ASC;
Response
┌────d─┬─count()─┬─round(avg(price), 2)─┬─bar(avg(price), 0, 50, 100)───────────┐
│ 1880 │ 2 │ 0.42 │ ▋ │
│ 1890 │ 7 │ 0.85 │ █▋ │
│ 1900 │ 399 │ 0.49 │ ▊ │
│ 1910 │ 589 │ 0.68 │ █▎ │
│ 1920 │ 280 │ 0.56 │ █ │
│ 1930 │ 74 │ 0.42 │ ▋ │
│ 1940 │ 119 │ 0.59 │ █▏ │
│ 1950 │ 134 │ 1.09 │ ██▏ │
│ 1960 │ 272 │ 0.92 │ █▋ │
│ 1970 │ 108 │ 1.18 │ ██▎ │
│ 1980 │ 88 │ 2.82 │ █████▋ │
│ 1990 │ 184 │ 3.68 │ ███████▎ │
│ 2000 │ 21 │ 7.14 │ ██████████████▎ │
│ 2010 │ 6 │ 18.42 │ ████████████████████████████████████▋ │
└──────┴─────────┴──────────────────────┴───────────────────────────────────────┘
伏特加
Query
SELECT
round(toUInt32OrZero(extract(menu_date, '^\\d{4}')), -1) AS d,
count(),
round(avg(price), 2),
bar(avg(price), 0, 50, 100)
FROM menu_item_denorm
WHERE (menu_currency IN ('Dollars', '')) AND (d > 0) AND (d < 2022) AND (dish_name ILIKE '%vodka%')
GROUP BY d
ORDER BY d ASC;
Response
┌────d─┬─count()─┬─round(avg(price), 2)─┬─bar(avg(price), 0, 50, 100)─┐
│ 1910 │ 2 │ 0 │ │
│ 1920 │ 1 │ 0.3 │ ▌ │
│ 1940 │ 21 │ 0.42 │ ▋ │
│ 1950 │ 14 │ 0.59 │ █▏ │
│ 1960 │ 113 │ 2.17 │ ████▎ │
│ 1970 │ 37 │ 0.68 │ █▎ │
│ 1980 │ 19 │ 2.55 │ █████ │
│ 1990 │ 86 │ 3.6 │ ███████▏ │
│ 2000 │ 2 │ 3.98 │ ███████▊ │
└──────┴─────────┴──────────────────────┴─────────────────────────────┘
ILIKE '%vodka%'，这写法绝对很有态度。
鱼子酱
Query
SELECT
round(toUInt32OrZero(extract(menu_date, '^\\d{4}')), -1) AS d,
count(),
round(avg(price), 2),
bar(avg(price), 0, 50, 100),
any(dish_name)
FROM menu_item_denorm
WHERE (menu_currency IN ('Dollars', '')) AND (d > 0) AND (d < 2022) AND (dish_name ILIKE '%caviar%')
GROUP BY d
ORDER BY d ASC;
Response
┌────d─┬─count()─┬─round(avg(price), 2)─┬─bar(avg(price), 0, 50, 100)──────┬─any(dish_name)──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ 1090 │ 1 │ 0 │ │ Caviar │
│ 1880 │ 3 │ 0 │ │ Caviar │
│ 1890 │ 39 │ 0.59 │ █▏ │ Butter and caviar │
│ 1900 │ 1014 │ 0.34 │ ▋ │ Anchovy Caviar on Toast │
│ 1910 │ 1588 │ 1.35 │ ██▋ │ 1/1 Brötchen Caviar │
│ 1920 │ 927 │ 1.37 │ ██▋ │ ASTRAKAN CAVIAR │
│ 1930 │ 289 │ 1.91 │ ███▋ │ Astrachan caviar │
│ 1940 │ 201 │ 0.83 │ █▋ │ (SPECIAL) Domestic Caviar Sandwich │
│ 1950 │ 81 │ 2.27 │ ████▌ │ Beluga Caviar │
│ 1960 │ 126 │ 2.21 │ ████▍ │ Beluga Caviar │
│ 1970 │ 105 │ 0.95 │ █▊ │ BELUGA MALOSSOL CAVIAR AMERICAN DRESSING │
│ 1980 │ 12 │ 7.22 │ ██████████████▍ │ Authentic Iranian Beluga Caviar the world's finest black caviar presented in ice garni and a sampling of chilled 100° Russian vodka │
│ 1990 │ 74 │ 14.42 │ ████████████████████████████▋ │ Avocado Salad, Fresh cut avocado with caviare │
│ 2000 │ 3 │ 7.82 │ ███████████████▋ │ Aufgeschlagenes Kartoffelsueppchen mit Forellencaviar │
│ 2010 │ 6 │ 15.58 │ ███████████████████████████████▏ │ "OYSTERS AND PEARLS" "Sabayon" of Pearl Tapioca with Island Creek Oysters and Russian Sevruga Caviar │
└──────┴─────────┴──────────────────────┴──────────────────────────────────┴─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘