I have just realised on my AWS Aurora postgres cluster having functions with temp_tables are not friendly with read replicas. I need to do a re-write (using CTEs) - anyway.... How do I take a json object with arrays nested and flatten them to a table like so:
{
"data": [
{
"groupName": "TeamA",
"groupCode": "12",
"subGroupCodes": [
"11"
]
},
{
"groupName": "TeamB",
"groupCode": "13",
"subGroupCodes": [
"15", "22"
]
}
]
}
I would like the output table to be:
groupName groupCode subGroupCodes
TeamA 12 11
TeamB 13 15
TeamB 13 22
I know I can get most of the way there with:
SELECT j."groupCode" as int, j."groupName" as pupilgroup_name
FROM json_to_recordset(p_in_filters->'data') j ("groupName" varchar(50), "groupCode" int)
But I just need to get the subGroupCodes as well but unpacking the array and joining to the correct parent groupCodes.
You need to first unnest the array, and then another unnest to get the subgroup codes:
with data (j) as (
values ('{
"data": [
{
"groupName": "TeamA",
"groupCode": "12",
"subGroupCodes": [
"11"
]
},
{
"groupName": "TeamB",
"groupCode": "13",
"subGroupCodes": [
"15", "22"
]
}
]
}'::jsonb)
)
select e ->> 'groupName' as group_name,
e ->> 'groupCode' as code,
sg.*
from data d
cross join lateral jsonb_array_elements(d.j -> 'data') as e(g)
cross join lateral jsonb_array_elements_text(g -> 'subGroupCodes') as sg(subgroup_code)
Related
My json format in one of the sql columns "jsoncol" in the table "jsontable" is like below.
Kindly help me to get this data using JSON_QUERY or JSON_VALUE
Please pay attention to the brackets and double quotes in the key value pairs...
{
"Company": [
{
"Info": {
"Address": "123"
},
"Name": "ABC",
"Id": 999
},
{
"Info": {
"Address": "456"
},
"Name": "XYZ",
"Id": 888
}
]
}
I am trying to retrieve all the company names using sql query. Thanks in advance
You can use:
SELECT j.name
FROM table_name t
CROSS APPLY JSON_TABLE(
t.value,
'$.Company[*]'
COLUMNS(
name VARCHAR2(200) PATH '$.Name'
)
) j
Which, for the sample data:
CREATE TABLE table_name (
value CLOB CHECK (value IS JSON)
);
INSERT INTO table_name (value)
VALUES ('{
"Company": [
{
"Info": {
"Address": "123"
},
"Name": "ABC",
"Id": 999
},
{
"Info": {
"Address": "456"
},
"Name": "XYZ",
"Id": 888
}
]
}');
Outputs:
NAME
ABC
XYZ
db<>fiddle here
You can easily use JSON_TABLE() function for this case rather provided the DB version is at least 12.1.0.2 such as
SELECT name
FROM jsontable,
JSON_TABLE(jsoncol,
'$' COLUMNS(NESTED PATH '$."Company"[*]'
COLUMNS(name VARCHAR2 PATH '$."Name"')))
Demo
I have documents in a bucket called blocks in the following format:
{
"random_field": 1,
"transactions": [{
"id": "CCCCC",
"inputs": [{
"tx_id": "AAAAA",
"index": 0
},{
"tx_id": "BBBBB",
"index": 1
}]
}]
}
{
"transactions": [{
"id": "AAAAA",
"outputs": [{
"field1": "value123",
"field2": "value456"
},{
"field1": "ignore",
"field2": "ignore"
}]
}]
}
{
"transactions": [{
"id": "BBBBB",
"outputs": [{
"field1": "ignored",
"field2": "ignored"
},{
"field1": "value999",
"field2": "value888"
}]
}]
}
and I need to map the inputs from the first document to the corresponding outputs of the second and third documents. The way to do it manually is to, for each input, find a transaction with id equal to the input's tx_id, and then get the item from the outputs array based on the index of the input. To exemplify, this is the object I would like to return in this scenario:
{
"random_field": 1,
"transactions": [{
"id": "CCCCC",
"inputs": [{
"tx_id": "AAAAA",
"index": 0,
"output": {
"field1": "value123",
"field2": "value456"
}
},{
"tx_id": "BBBBB",
"index": 1,
"output": {
"field1": "value999",
"field2": "value888"
}
}]
}]
}
I managed to come up with the following query:
SELECT b.random_field,
b.transactions -- how to map this?
FROM blocks b
UNNEST b.transactions t
UNNEST t.inputs input
JOIN blocks `source` ON (ANY tx IN `source`.transactions SATISFIES tx.`id` = input.tx_id END)
UNNEST `source`.transactions source_tx
UNNEST source_tx.outputs o
WHERE (ANY tx IN b.transactions SATISFIES tx.`id` = 'AAAAA' END) LIMIT 1;
I suppose there should be a way to map b.transactions.inputs by using source_tx.outputs, but I couldn't find how.
I came across this other answer, but I don't really understand how it applies to my scenario. Maybe it does, but I am very new to Couchbase, so I am very much lost: How to map array values in one document to another and display in result
Basically you want inline some other document into current document using condition.
Instead of JOINs+ GROUPS use subquery expressions + correlated subqueries. (b.*, "abc" AS transactions, selects all fields of b and adds transactions (if already exist overwrite else adds)
CREATE INDEX ix1 ON blocks (ALL ARRAY FOR ot.id FOR ot IN transactions END);
SELECT b.*,
(SELECT t.*,
(SELECT i.*,
(SELECT RAW ot
FROM blocks AS o
UNNEST o.transactions AS ot
UNNEST ot.outputs AS oto
WHERE i.tx_id = ot.id AND i.`index` = UNNEST_POS(oto))[0] AS output
FROM t.`inputs` AS i) AS inputs
FROM b.transactions AS t) AS transactions
FROM blocks AS b
WHERE ANY tx IN b.transactions SATISFIES tx.`inputs` IS NOT NULL END ;
OR
SELECT b.*,
(SELECT t.*,
(SELECT i.*,
(SELECT RAW ot.outputs[i.`index`]
FROM blocks AS o
UNNEST o.transactions AS ot
WHERE i.tx_id = ot.id
LIMIT 1)[0] AS output
FROM t.`inputs` AS i) AS inputs
FROM b.transactions AS t) AS transactions
FROM blocks AS b
WHERE ANY tx IN b.transactions SATISFIES tx.`inputs` IS NOT NULL END ;
I have a table (log_table) and in this table there is a nested array json field (activities). With using this activities field, I want to normalize my row.
log_table:
- id:long
- activities:json
- date:timestamp
example activities field:
[
{
"actionType":"NOTIFICATION",
"items":null
},
{
"actionType":"MUTATION",
"items":[
{
"id":387015007,
"name":"epic",
"value":{
"currency":"USD",
"amount":1.76
}
},
{
"id":386521039,
"name":"test",
"value":{
"currency":"USD",
"amount":1.76
}
}
]
}
]
As query, I've tried:
select
*
from
log_table l,
json_array_elements(l.activities) elems,
json_array_elements(elems->'items') obj;
With this query, I got error like below:
ERROR: cannot call json_array_elements on a scalar
Is there any suggestion?
The lack of items should be marked as [null], not null. You can use the case expression to correct this, e.g.:
select elems->>'actionType' as action_type, obj
from log_table
cross join jsonb_array_elements(l.activities::jsonb) elems
cross join jsonb_array_elements(case elems->'items' when 'null' then '[null]' else elems->'items' end) obj
action_type | obj
--------------+---------------------------------------------------------------------------------
NOTIFICATION | null
MUTATION | {"id": 387015007, "name": "epic", "value": {"amount": 1.76, "currency": "USD"}}
MUTATION | {"id": 386521039, "name": "test", "value": {"amount": 1.76, "currency": "USD"}}
(3 rows)
I have a table with a column that holds valid JSON, heres an example of the JSON structure:
{
"Requirements": {
"$values": [
{
"$type": "List",
"ListId": "956cf9c5-24ab-47d9-8082-940118f2f1a3",
"DefaultValue": "",
"MultiSelect": true,
"Selected": null,
"MultiSelected": {
"$type": "ListItem",
"$values": [
"Value1",
"Value2",
"Value3"
]
}
},
{
"$type": "List",
"ListId": "D11149DD-A682-4BC7-A87D-567954779234",
"DefaultValue": "",
"MultiSelect": true,
"Selected": null,
"MultiSelected": {
"$type": "ListItem",
"$values": [
"Value4",
"Value5",
"Value6",
"Value7"
]
}
}
]
}
}
I need to return the values from MultiSelected collection depending on the value of ListID.
I'm using the following JSON Path to retun value
$.Requirements."$values"[?(#.ListId=='956cf9c5-24ab-47d9-8082-940118f2f1a3')].MultiSelected."$values"
This worked fine in a JSON Expression tester.
But when I try to use it to query the table I get the following error:
JSON path is not properly formatted. Unexpected character '?' is found at position 25.
The query I'm using is as follows:
SELECT ID AS PayloadID,
Items.Item AS ItemsValues
FROM dbo.Payload
CROSS APPLY ( SELECT *
FROM OPENJSON( JSON_QUERY( Payload, '$.Requirements."$values"[?(#.ListId==''956cf9c5-24ab-47d9-8082-940118f2f1a3'')].MultiSelected."$values"' ) )
WITH ( Item nvarchar(200) '$' ) ) AS Items
WHERE ID = 3
I've tried replacing
?(#.ListId==''956cf9c5-24ab-47d9-8082-940118f2f1a3'')
with 0 and it works fine on SQL Server.
My question is, is filter syntax ?(...) supported in JSON_QUERY or is there something else I should be doing?
The database is running on Azure, were the database compatability level is set to SQL Server 2017 (140).
Thanks for your help in advance.
Andy
I would use openjson twice in stead
drop table if exists #payload
create table #payload(ID int,Payload nvarchar(max))
insert into #payload VALUES
(3,N'
{
"Requirements": {
"$values": [
{
"$type": "List",
"ListId": "956cf9c5-24ab-47d9-8082-940118f2f1a3",
"DefaultValue": "",
"MultiSelect": true,
"Selected": null,
"MultiSelected": {
"$type": "ListItem",
"$values": [
"Value1",
"Value2",
"Value3"
]
}
},
{
"$type": "List",
"ListId": "D11149DD-A682-4BC7-A87D-567954779234",
"DefaultValue": "",
"MultiSelect": true,
"Selected": null,
"MultiSelected": {
"$type": "ListItem",
"$values": [
"Value4",
"Value5",
"Value6",
"Value7"
]
}
}
]
}
}'
)
SELECT ID AS PayloadID,
Items.[value]
FROM #Payload a
CROSS APPLY OPENJSON( Payload, '$.Requirements."$values"' ) with ( ListId varchar(50),MultiSelected nvarchar(max) as json) b
CROSS APPLY OPENJSON( MultiSelected,'$."$values"' ) Items
where
a.id=3
AND b.listid='956cf9c5-24ab-47d9-8082-940118f2f1a3'
Reply:
+-----------+--------+
| PayloadID | value |
+-----------+--------+
| 3 | Value1 |
| 3 | Value2 |
| 3 | Value3 |
+-----------+--------+
Document sample from my giata_properties bucket: link
Relevant json paste
{
"propertyCodes": {
"provider": [
{
"code": [
{
"value": [
{
"value": "304387"
}
]
}
],
"providerCode": "hotelbeds",
"providerType": "gds"
},
{
"code": [
{
"value": [
{
"name": "Country Code",
"value": "EG"
},
{
"name": "City Code",
"value": "HRG"
},
{
"name": "Hotel Code",
"value": "91U"
}
]
}
],
"providerCode": "gta",
"providerType": "gds"
}
]
},
"name": "Arabia Azur Resort"
}
I want a query (and an index) to retrieve a document based on propertyCodes.provider.code.value.value and propertyCodes.provider.providerCode. I've managed to do each separately but I'm not sure how to merge both of them in a single query.
SELECT meta().id FROM giata_properties AS gp USE INDEX(`#primary`) WHERE ANY v WITHIN gp.propertyCodes.provider[*].code SATISFIES v.`value` = '150613' END;
SELECT meta().id FROM giata_properties AS gp USE INDEX(`#primary`) WHERE ANY v within gp.propertyCodes.provider[*].providerCode SATISFIES v = 'hotelbeds' END;
So for example I want to fetch the document that includes propertyCodes.provider.code.value.value of 304387 and that provider is also hotelbeds, because code value can be duplicated over documents, but code and providerCode combination is unique.
Here are the query and the indexes.
The query.
SELECT META().id
FROM giata_properties AS gp
WHERE ANY p IN propertyCodes.provider SATISFIES ( ANY v WITHIN p.code SATISFIES v.`value` = '304387' END ) AND p.providerCode = 'hotelbeds' END;
The indexes.
CREATE INDEX idx_value ON giata_properties
( DISTINCT ARRAY ( DISTINCT ARRAY v.`value` FOR v WITHIN p.code END ) FOR p IN propertyCodes.provider END );
CREATE INDEX idx_providerCode ON giata_properties
( DISTINCT ARRAY p.providerCode FOR p IN propertyCodes.provider END );