在 IF 语句中嵌套展平；重新键入多层展平
Posted
技术标签:
【中文标题】在 IF 语句中嵌套展平；重新键入多层展平【英文标题】：Nested flatten in IF statement; re-keying in multi layer flatten 【发布时间】：2020-09-13 14:56:02 【问题描述】：当前查询：
with
report as (
select * from Report_Data
where type = 'pnl'
and (created_at > '2020-01-01' or updated_at > '2020-01-01')
),
flattened as (
select
report.company_ID as "CID",
report.type,
report.from_date as "Report_Date",
f.*
--f.value:"title"
from report,
lateral flatten(input =>prepared:profitAndLoss:report:sections) f
--,lateral flatten(input =>f.value)f1
)
select * from flattened where CID = 908 limit 100;
f 的电流输出：
[
"rows": [
"emphasize": false,
"name": "Product Sales",
"values": [
"rawValue": 19889.16,
"value": "19,889"
,
"rawValue": 56181.94,
"value": "56,182"
]
,
"emphasize": false,
"name": "Credits from Vendors",
"values": [
"rawValue": 38,
"value": "38"
,
"rawValue": 47.16,
"value": "47"
]
,
"emphasize": false,
"name": "Sales - PayPal",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 275,
"value": "275"
]
,
"emphasize": false,
"name": "Sales Returns and Allowances",
"values": [
"rawValue": -379.28,
"value": "(379)"
,
"rawValue": -941.78,
"value": "(942)"
]
],
"summary":
"emphasize": false,
"name": "Total Revenue",
"values": [
"rawValue": 19547.88,
"value": "19,548"
,
"rawValue": 55562.32,
"value": "55,562"
]
,
"title": "Revenue"
,
"rows": [
"emphasize": false,
"name": "Product Costs",
"values": [
"rawValue": 9857.22,
"value": "9,857"
,
"rawValue": 31907.32,
"value": "31,907"
]
,
"emphasize": false,
"name": "Supplies",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 109.92,
"value": "110"
]
],
"summary":
"emphasize": false,
"name": "Total Cost of Sales",
"values": [
"rawValue": 9857.22,
"value": "9,857"
,
"rawValue": 32017.24,
"value": "32,017"
]
,
"title": "Less Cost of Sales"
,
"rows": [
"emphasize": true,
"name": "Gross Profit",
"values": [
"rawValue": 9690.66,
"value": "9,691"
,
"rawValue": 23545.08,
"value": "23,545"
]
],
"summary":
"name": "",
"values": []
,
"title": ""
,
"rows": [
"emphasize": false,
"name": "Advertising & Marketing",
"values": [
"rawValue": 100,
"value": "100"
,
"rawValue": 100,
"value": "100"
]
,
"emphasize": false,
"name": "Advertising Expenses",
"values": [
"rawValue": 22,
"value": "22"
,
"rawValue": 689,
"value": "689"
]
,
"emphasize": false,
"name": "Bank Fees",
"values": [
"rawValue": 53,
"value": "53"
,
"rawValue": 53,
"value": "53"
]
,
"emphasize": false,
"name": "Business Telephone and Internet",
"values": [
"rawValue": 10.08,
"value": "10"
,
"rawValue": 30.24,
"value": "30"
]
,
"emphasize": false,
"name": "Charitable Contributions",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": -81.89,
"value": "(82)"
]
,
"emphasize": false,
"name": "Computer and Software Expense",
"values": [
"rawValue": 20,
"value": "20"
,
"rawValue": 20,
"value": "20"
]
,
"emphasize": false,
"name": "Dues and Subscriptions",
"values": [
"rawValue": 0.99,
"value": "1"
,
"rawValue": 135.99,
"value": "136"
]
,
"emphasize": false,
"name": "Fees and Licenses",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 50,
"value": "50"
]
,
"emphasize": false,
"name": "Insurance Expense - General Liability",
"values": [
"rawValue": 101.6,
"value": "102"
,
"rawValue": 304.8,
"value": "305"
]
,
"emphasize": false,
"name": "Legal & Professional",
"values": [
"rawValue": 200,
"value": "200"
,
"rawValue": 600,
"value": "600"
]
,
"emphasize": false,
"name": "Merchant Account Fees",
"values": [
"rawValue": 444.91,
"value": "445"
,
"rawValue": 1221.71,
"value": "1,222"
]
,
"emphasize": false,
"name": "Office Expenses",
"values": [
"rawValue": 169.66,
"value": "170"
,
"rawValue": 988.42,
"value": "988"
]
,
"emphasize": false,
"name": "Payroll Processing Fee",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 0.45,
"value": "0"
]
,
"emphasize": false,
"name": "Rent",
"values": [
"rawValue": 1500,
"value": "1,500"
,
"rawValue": 6000,
"value": "6,000"
]
,
"emphasize": false,
"name": "Taxes",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 3000,
"value": "3,000"
]
,
"emphasize": false,
"name": "Travel",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 194.35,
"value": "194"
]
,
"emphasize": false,
"name": "Utilities",
"values": [
"rawValue": 559.59,
"value": "560"
,
"rawValue": 1769.63,
"value": "1,770"
]
],
"summary":
"emphasize": false,
"name": "Total Operating Expenses",
"values": [
"rawValue": 3181.83,
"value": "3,182"
,
"rawValue": 15075.7,
"value": "15,076"
]
,
"title": "Operating Expenses"
,
"rows": [
"emphasize": false,
"name": "Operating Income / (Loss)",
"values": [
"rawValue": 6508.83,
"value": "6,509"
,
"rawValue": 8469.38,
"value": "8,469"
]
],
"summary":
"name": "",
"values": []
,
"title": ""
,
"rows": [
"emphasize": false,
"name": "Interest",
"values": [
"rawValue": -85.44,
"value": "(85)"
,
"rawValue": -434.74,
"value": "(435)"
]
,
"emphasize": false,
"name": "Interest Income",
"values": [
"rawValue": 0.57,
"value": "1"
,
"rawValue": 1.24,
"value": "1"
]
,
"emphasize": false,
"name": "Other Income",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 100,
"value": "100"
]
],
"summary":
"emphasize": false,
"name": "Total Other Income and Expense",
"values": [
"rawValue": -84.87,
"value": "(85)"
,
"rawValue": -333.5,
"value": "(334)"
]
,
"title": "Other Income and Expense"
,
"rows": [
"emphasize": true,
"name": "Net Income / (Loss) before Tax",
"values": [
"rawValue": 6423.96,
"value": "6,424"
,
"rawValue": 8135.88,
"value": "8,136"
]
],
"summary":
"name": "",
"values": []
,
"title": ""
,
"rows": [
"emphasize": false,
"name": "Net Income",
"values": [
"rawValue": 6423.96,
"value": "6,424"
,
"rawValue": 8135.88,
"value": "8,136"
]
],
"summary":
"name": "",
"values": []
,
"title": ""
,
"rows": [
"emphasize": false,
"name": "Total Comprehensive Income",
"values": [
"rawValue": 6423.96,
"value": "6,424"
,
"rawValue": 8135.88,
"value": "8,136"
]
],
"summary":
"name": "",
"values": []
,
"title": ""
]
https://codebeautify.org/jsonviewer
将 c/p JSON 推荐到上面的链接中并选择“树查看器”以跟随我的导航参考
问题：
雪花/ASNI SQL：
对于数组中的每个索引，我需要评估一个检查“title”键的逻辑函数，如果“title”键的值为 X（为了本练习，我们选择“Revenue”），返回“行”键的每个索引的“名称”和“值”。在 rows:values 下，您有 2 个值。它们需要重新键入 MTD 和 YTD。它们始终存储为 [0]:MTD 和 [1]:YTD。我希望这些新键出现在我命名为“周期性”的列中。我知道这只是雪花中一个棘手的扁平化问题，但我整个周末都在多次不同的迭代中苦苦挣扎，想看看是否有人可以试一试。感谢您的帮助！
期望的输出： CID, TYPE, REPORT_DATE, "Title" Key, rows:Name as "Name", 周期性, rows:values:[0]:rawValue
编辑：带有值的示例输出列：
CID TYPE REPORT_DATE "Title" Key rows:Name as "Name"
47 PNL 2020-01-01 REVENUE "Sales - PayPal"
periodicity rows:values:[0]:rawValue
"MTD" 0
【问题讨论】：
如果您的示例 json 只有“一个”块来关心“收入”，我可以建议您需要另一个不是“收入”的块来展示数据的混合性质。此外，虽然您已经在输出中描述了所需列的名称，但您还没有将您希望在示例中看到的值放入示例中。这将允许人们弥合您的描述和您想要的结果之间的差距。 您是否建议prepared:profitAndLoss:report:sections 是附加的 JSON，或者实际上 f.* 的结果是 JSON，因为 f.* 的输出看起来不像您的 JSON，这让我想知道什么你认为你的 JSON 试图代表。
已编辑以显示示例输出，您成功了，谢谢。其他块在其他“title”键下，例如“title”：“Operating Expenses”。至于prepared:profitAndLoss:report:sections，这不是我能够将JSON从原始范围缩小到的。示例 JSON 是需要展平的。
所以如果 json 已经“正确”展平，还有什么问题吗？
不！你明白了！
【参考方案1】：
所以压扁不会太难。如果你一步一步地走：
with random_json as (
select '2020-05-27'::date as report_date, parse_json('[
"rows": [
"emphasize": false,
"name": "Product Sales",
"values": [
"rawValue": 19889.16,
"value": "19,889"
,
"rawValue": 56181.94,
"value": "56,182"
]
,
"emphasize": false,
"name": "Credits from Vendors",
"values": [
"rawValue": 38,
"value": "38"
,
"rawValue": 47.16,
"value": "47"
]
,
"emphasize": false,
"name": "Sales - PayPal",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 275,
"value": "275"
]
,
"emphasize": false,
"name": "Sales Returns and Allowances",
"values": [
"rawValue": -379.28,
"value": "(379)"
,
"rawValue": -941.78,
"value": "(942)"
]
],
"summary":
"emphasize": false,
"name": "Total Revenue",
"values": [
"rawValue": 19547.88,
"value": "19,548"
,
"rawValue": 55562.32,
"value": "55,562"
]
,
"title": "Revenue"
,
"rows": [
"emphasize": false,
"name": "Product Costs",
"values": [
"rawValue": 9857.22,
"value": "9,857"
,
"rawValue": 31907.32,
"value": "31,907"
]
,
"emphasize": false,
"name": "Supplies",
"values": [
"rawValue": 0,
"value": "0"
,
"rawValue": 109.92,
"value": "110"
]
],
"summary":
"emphasize": false,
"name": "Total Cost of Sales",
"values": [
"rawValue": 9857.22,
"value": "9,857"
,
"rawValue": 32017.24,
"value": "32,017"
]
,
"title": "Less Cost of Sales"
]') as json
), flatten_a AS (
select
r.report_date
,f.value:rows as r_rows
,f.value:summary as r_summary
,f.value:title as r_title
from random_json r
,lateral flatten(input=>r.json) f
), flatten_b AS (
select a.*
, f.value:name as rr_name
, f.value:values as rr_values
from flatten_a a
,lateral flatten(input=>a.r_rows) f
)
select b.report_date
,b.r_title as title_key
,b.rr_name as name
,case when f.index =0 then 'MTD' when f.index=1 then 'YTD' else 'WTF' end as periodicity
, f.value:rawValue raw_value
from flatten_b b
,lateral flatten(input=>b.rr_values) f;
给予：
REPORT_DATE TITLE_KEY NAME PERIODICITY RAW_VALUE
2020-05-27 "Revenue" "Product Sales" MTD 19889.16
2020-05-27 "Revenue" "Product Sales" YTD 56181.94
2020-05-27 "Revenue" "Credits from Vendors" MTD 38
2020-05-27 "Revenue" "Credits from Vendors" YTD 47.16
2020-05-27 "Revenue" "Sales - PayPal" MTD 0
2020-05-27 "Revenue" "Sales - PayPal" YTD 275
2020-05-27 "Revenue" "Sales Returns and Allowances" MTD -379.28
2020-05-27 "Revenue" "Sales Returns and Allowances" YTD -941.78
2020-05-27 "Less Cost of Sales" "Product Costs" MTD 9857.22
2020-05-27 "Less Cost of Sales" "Product Costs" YTD 31907.32
2020-05-27 "Less Cost of Sales" "Supplies" MTD 0
2020-05-27 "Less Cost of Sales" "Supplies" YTD 109.92
如果您在第一句中提到的仅希望上述“收入”行使结果集也成为 CTE，并左加入您的“报告”和 ON 过滤器以获取收入。
【讨论】：
以上是关于在 IF 语句中嵌套展平；重新键入多层展平的主要内容，如果未能解决你的问题，请参考以下文章