Uber Eats and GRUBHUB analysis

1. Running the first command to retrieve the data on Big Query to get data.

SELECT * FROM arboreal-vision-339901.take_home_v2.virtual_kitchen_ubereats_hours LIMIT

1000;
SELECT * FROM arboreal-vision-339901.take_home_v2.virtual_kitchen_grubhub_hours LIMIT
1000;

2. Deleting duplicate rows by keeping only the most recent timestamps.

SELECT t1.response, t1.timestamp, t1.slug, t1.b_name, t1.vb_name, t1.vb_platform,

t1.vb_address, t1.token
FROM `ordinal-link-390505.dataset.grubhub_initial` AS t1
INNER JOIN (
SELECT slug, MAX(timestamp) AS max_timestamp
FROM `ordinal-link-390505.dataset.grubhub_initial`
GROUP BY slug ) t2
ON t1.slug = t2.slug AND t1.timestamp = t2.max_timestamp;

SELECT u1.response, u1.timestamp, u1.slug, u1.b_name, u1.vb_name, u1.vb_platform,

u1.vb_address, u1.token
FROM `ordinal-link-390505.dataset.ubereats_initial` AS u1
INNER JOIN (
SELECT slug, MAX(timestamp) AS max_timestamp
FROM `ordinal-link-390505.dataset.ubereats_initial`
GROUP BY slug) u2
ON u1.slug = u2.slug AND u1.timestamp = u2.max_timestamp;
3. Extracting the required JSON Data from the grubhub_initial_1 table.
CREATE TEMP FUNCTION JSON_Parser(jsonString STRING)
RETURNS STRUCT<days_of_week ARRAY<STRING>, from_time STRING, to_time STRING>
LANGUAGE js AS """
let parsed = JSON.parse(jsonString);
return {
days_of_week: parsed.days_of_week,
from_time: parsed.from,
to_time: parsed.to
};
""";
WITH Table1 AS
(SELECT tree, slug, CONCAT(b_name, ' ', vb_name) AS Business
FROM `ordinal-link-390505.dataset.grubhub_initial_2` AS gh
CROSS JOIN UNNEST(JSON_QUERY_ARRAY(gh.response,
'$.availability_by_catalog.STANDARD_DELIVERY.schedule_rules')) AS tree)
SELECT JSON_Parser(TO_JSON_STRING(tree)).*, Business
FROM Table1;

4. The JSON path for the UberEATS JSON object contains a variable. Since JSON path query cannot be
dynamic, we will drill down on the node we require up until the variant.
CREATE TEMP FUNCTION EXTRACT_KV_PAIRS(json_str STRING)
RETURNS ARRAY<STRUCT<key STRING, value STRING>>
LANGUAGE js AS """
try{
const json_dict = JSON.parse(json_str);
const all_kv = Object.entries(json_dict).map(
(r)=>Object.fromEntries([["key", r[0]],["value",
JSON.stringify(r[1])]]));
return all_kv;
} catch(e) { return [{"key": "error","value": e}];}
""";
SELECT EXTRACT_KV_PAIRS(TO_JSON_STRING(response.data.menus)), slug from
`silicon-alpha-390017.loop_assignment.ubereats`;
5. Extracting the day of the week, opening, and closing times of the restaurants registered with
Ubereats.

WITH regularHours AS (
SELECT JSON_QUERY_ARRAY(ujk.f0_[SAFE_OFFSET(0)].value, '$.sections.0.regularHours')
AS regularHoursArray, ujk.slug, CONCAT(ue.b_name, ' ', ue.vb_name) AS name
FROM `ordinal-link-390505.dataset.ubereats_initial_3_var` AS ujk
JOIN `ordinal-link-390505.dataset.ubereats_initial_2` AS ue
ON ujk.slug = ue.slug
),
regularHoursUnnested AS (
SELECT
JSON_QUERY(regularHour, '$.daysBitArray') AS daysBitArray,
JSON_EXTRACT_SCALAR(regularHour, '$.endTime') AS endTime,
JSON_EXTRACT_SCALAR(regularHour, '$.startTime') AS startTime,
slug, name
FROM regularHours
CROSS JOIN UNNEST(regularHoursArray) AS regularHour
)
SELECT
(CASE
WHEN daysBitArray = '[true,false,false,false,false,false,false]' THEN 'Sunday'
WHEN daysBitArray = '[false,true,false,false,false,false,false]' THEN 'Monday'
WHEN daysBitArray = '[false,false,true,false,false,false,false]' THEN 'Tuesday'
WHEN daysBitArray = '[false,false,false,true,false,false,false]' THEN 'Wednesday'
WHEN daysBitArray = '[false,false,false,false,true,false,false]' THEN 'Thursday'
WHEN daysBitArray = '[false,false,false,false,false,true,false]' THEN 'Friday'
WHEN daysBitArray = '[false,false,false,false,false,false,true]' THEN 'Saturday'
ELSE 'Unknown'
END) AS dayOfWeek, startTime, endTime, name, slug
FROM regularHoursUnnested
ORDER BY name, (CASE dayOfWeek
WHEN 'Monday' THEN 1
WHEN 'Tuesday' THEN 2
WHEN 'Wednesday' THEN 3
WHEN 'Thursday' THEN 4
WHEN 'Friday' THEN 5
WHEN 'Saturday' THEN 6
WHEN 'Sunday' THEN 7
ELSE 8 END)
6. Unnesting the day of the week column in the Grubhub table so it can be analyzed further.

CREATE TABLE `ordinal-link-390505.dataset.grubhub_initial_4` AS

SELECT
(SELECT * FROM UNNEST(days_of_week) LIMIT 1) AS days_of_week,
from_time,
to_time,
Business,
slug
FROM `dataset.grubhub_initial_3`
7. Converting the startTime and endTime columns in the Ubereats and Grubhub table from dtype string
to dtype time and creating a new table.

CREATE TABLE `ordinal-link-390505.dataset.ubereats_initial_4_time` AS

SELECT
dayOfWeek,
PARSE_TIME('%R', startTime) AS startTime,
PARSE_TIME('%R', endTime) AS endTime,
name,
slug
FROM `ordinal-link-390505.dataset.ubereats_initial_4`
ORDER BY name;

CREATE TABLE `ordinal-link-390505.dataset.grubhub_initial_4_time` AS

SELECT
days_of_week,
PARSE_TIME('%T', SUBSTR(from_time, 1, 8)) AS from_time,
PARSE_TIME('%T', SUBSTR(to_time, 1, 8)) AS to_time,
Business AS name,
slug
FROM `ordinal-link-390505.dataset.grubhub_initial_4`
ORDER BY name
8. Converting day of week column in Ubereats and Grubhub tables to lowercase.

CREATE TABLE `ordinal-link-390505.dataset.grubhub_initial_5` AS

SELECT
LOWER(days_of_week) AS dayOfWeek,
from_time,
to_time,
name,
slug
FROM `ordinal-link-390505.dataset.grubhub_initial_4_time`

CREATE TABLE `ordinal-link-390505.dataset.ubereats_initial_5` AS

SELECT
LOWER(dayOfWeek) AS dayOfWeek,
startTime AS from_time,
endTime AS to_time,
name,
slug
FROM `ordinal-link-390505.dataset.ubereats_initial_4_time`
9. Creating a table with matched Business names and a time difference between the opening and
closing times calculated in minutes.

SELECT
gh.slug AS Grubhub_Slug,gh.from_time, gh.to_time,
TIME_DIFF(gh.to_time, gh.from_time, MINUTE) +
(CASE
WHEN gh.to_time = TIME '00:00:00' THEN 1440
WHEN gh.from_time > TIME '12:00:00' AND gh.from_time > gh.to_time THEN 1440
WHEN gh.from_time > gh.to_time THEN 720
ELSE 0
END) AS GH_operating,
ue.slug AS Ubereats_Slug, ue.from_time, ue.to_time,
TIME_DIFF(ue.to_time, ue.from_time, MINUTE) +
(CASE
WHEN ue.to_time = TIME '00:00:00' THEN 1440
WHEN ue.from_time > TIME '12:00:00' AND ue.from_time > ue.to_time THEN 1440
WHEN ue.from_time > ue.to_time THEN 720
ELSE 0
END) AS UE_operating
FROM `ordinal-link-390505.dataset.ubereats_initial_5` AS ue
INNER JOIN `ordinal-link-390505.dataset.grubhub_initial_5` AS gh
ON ue.name = gh.name AND ue.dayOfWeek = gh.dayOfWeek
ORDER by ue.name;
10. Creating the final table in accordance with the desired output of the assignment.

SELECT Ubereats_Slug, UE_operating, Grubhub_Slug, GH_operating,

(CASE
WHEN from_time_1 = from_time AND to_time_1 = to_time THEN 'In Range'
WHEN (from_time_1 = from_time OR to_time_1 = to_time) AND ABS(GH_operating -
UE_operating) < 5 THEN 'Out of range with 5 mins difference'
WHEN (from_time_1 != from_time OR to_time_1 != to_time) AND ABS(GH_operating -
UE_operating) >= 5 THEN 'Out of Range'
ELSE NULL
END) AS In_Range
FROM `ordinal-link-390505.dataset.final_table`

Notes: -

Some businesses in the Uber Eats dataset had timings for days labelled ‘Unknown’. Such rows
were dropped automatically when Inner Joining. Similarly, some businesses had multiple
entries for the same days of the week. In this case, the first entry for the weekday was selected.