JSON SQL column in azure data factory - sql

Question

I have a JSON type SQL column in SQL table as below example. I want the below code to be converted into separate columns such as drugs as table name and other attribute as column name, how can I use adf or any other means please guide. The below code is a single column in a table called report where I need to convert this into separate columns .
{
"drugs": {
"Codeine": {
"bin": "Y",
"name": "Codeine",
"icons": [
93,
103
],
"drug_id": 36,
"pathway": {
"code": "prodrug",
"text": "is **inactive**, its metabolites are active."
},
"targets": [],
"rxnorm_id": "2670",
"priclasses": [
"Analgesic/Anesthesiology"
],
"references": [
1,
16,
17,
100
],
"subclasses": [
"Analgesic agent",
"Antitussive agent",
"Opioid agonist",
"Phenanthrene "
],
"metabolizers": [
"CYP2D6"
],
"phenotype_ids": {
"metabolic": "5"
},
"relevant_genes": [
"CYP2D6"
],
"dosing_guidelines": [
{
"text": "Reduced morphine formation. Use label recommended age- or weight-specific dosing. If no response, consider alternative analgesics such as morphine or a non-opioid.",
"source": "CPIC",
"guidelines_id": 1
},
{
"text": "Analgesia: select alternative drug (e.g., acetaminophen, NSAID, morphine-not tramadol or oxycodone) or be alert to symptoms of insufficient pain relief.",
"source": "DPWG",
"guidelines_id": 362
}
],
"drug_report_notes": [
{
"text": "Predicted codeine metabolism is reduced.",
"icons_id": 58,
"sort_key": 58,
"references_id": null
},
{
"text": "Genotype suggests a possible decrease in exposure to the active metabolite(s) of codeine.",
"icons_id": 93,
"sort_key": 56,
"references_id": null
},
{
"text": "Professional guidelines exist for the use of codeine in patients with this genotype and/or phenotype.",
"icons_id": 103,
"sort_key": 50,
"references_id": null
}
]
}

Answer 1

Since this json is already in a SQL column, you don't need ADF to break it down to parts. You can use JSON functions in SQL server to do that.
example of few first columns:
declare #json varchar(max) = '{
"drugs": {
"Codeine": {
"bin": "Y",
"name": "Codeine",
"icons": [
93,
103
],
"drug_id": 36,
"pathway": {
"code": "prodrug",
"text": "is **inactive**, its metabolites are active."
},
"targets": [],
"rxnorm_id": "2670",
"priclasses": [
"Analgesic/Anesthesiology"
],
"references": [
1,
16,
17,
100
],
"subclasses": [
"Analgesic agent",
"Antitussive agent",
"Opioid agonist",
"Phenanthrene "
],
"metabolizers": [
"CYP2D6"
],
"phenotype_ids": {
"metabolic": "5"
},
"relevant_genes": [
"CYP2D6"
],
"dosing_guidelines": [
{
"text": "Reduced morphine formation. Use label recommended age- or weight-specific dosing. If no response, consider alternative analgesics such as morphine or a non-opioid.",
"source": "CPIC",
"guidelines_id": 1
},
{
"text": "Analgesia: select alternative drug (e.g., acetaminophen, NSAID, morphine-not tramadol or oxycodone) or be alert to symptoms of insufficient pain relief.",
"source": "DPWG",
"guidelines_id": 362
}
],
"drug_report_notes": [
{
"text": "Predicted codeine metabolism is reduced.",
"icons_id": 58,
"sort_key": 58,
"references_id": null
},
{
"text": "Genotype suggests a possible decrease in exposure to the active metabolite(s) of codeine.",
"icons_id": 93,
"sort_key": 56,
"references_id": null
},
{
"text": "Professional guidelines exist for the use of codeine in patients with this genotype and/or phenotype.",
"icons_id": 103,
"sort_key": 50,
"references_id": null
}
]
}
}
}
select JSON_VALUE(JSON_QUERY(#json,'$.drugs.Codeine'),'$.bin') as bin,
JSON_VALUE(JSON_QUERY(#json,'$.drugs.Codeine'),'$.name') as name,
JSON_VALUE(JSON_QUERY(#json,'$.drugs.Codeine'),'$.drug_id') as drug_id,
JSON_VALUE(JSON_QUERY(#json,'$.drugs.Codeine'),'$.icons[0]') as icon_1
'
You need to decide how to handle arrays, such as icons, where there are multiple values inside the same element.
References:
JSON_QUERY function
JSON_VALUE function