How can I transform the JSON in app logic? - azure

Question

I got to as far as a json result from log analytic query API HTTP action call:
{
"tables": [
{
"name": "PrimaryResult",
"columns": [
{
"name": "TimeGenerated",
"type": "datetime"
},
{
"name": "_queue",
"type": "string"
},
{
"name": "_messages",
"type": "real"
}
],
"rows": [
[
"2022-06-03T03:20:00Z",
"queue1",
8073
],
[
"2022-06-03T03:20:00Z",
"queue2",
570
]
]
}
]
}
I need transform it to this following format, which essentially an adaptivecards.io card
{
"type": "ColumnSet",
"columns": [
{
"type": "Column",
"items": [
{
"type": "TextBlock",
"text": "**TimeGenerated**" //Columns[0].Name
},
{
"type": "TextBlock",
"text": "2022-06-03T03:20:00Z" //Rows[0][0]
},
{
"type": "TextBlock",
"text": "2022-06-03T03:20:00Z" //Rows[1][0]
}
]
},
{
"type": "Column",
"items": [
{
"type": "TextBlock",
"text": "**_queue**" //Columns[1].Name
},
{
"type": "TextBlock",
"text": "queue1" //Rows[0][1]
},
{
"type": "TextBlock",
"text": "queue2" //Rows[1][1]
}
]
},
{
"type": "Column",
"items": [
{
"type": "TextBlock",
"text": "**_messages**"
},
{
"type": "TextBlock",
"text": "8073"
},
{
"type": "TextBlock",
"text": "570"
}
]
}
]
}
How can I do this? I tried with nested for-loop action, however I'm stuck in composing the final variable that can be inserted in the right place.
It would be ideal to hold a transformed variable ahead of time so that I can just include it in the later stage as POST to another webhook.
Edit: Whilst the column is static, the number rows are dynamic.

Answer 1

Create a LogicApp and load in this definition to see how to do it ...
{
"definition": {
"$schema": "https://schema.management.azure.com/providers/Microsoft.Logic/schemas/2016-06-01/workflowdefinition.json#",
"actions": {
"Initialize_AdaptiveCard_JSON": {
"inputs": {
"variables": [
{
"name": "XML",
"type": "string",
"value": "{\n \"type\": \"ColumnSet\",\n \"columns\": [\n {\n \"type\": \"Column\",\n \"items\": [\n {\n \"type\": \"TextBlock\",\n \"text\": \"**#{variables('Object Variable')['tables'][0]['columns'][0]['Name']}**\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][0][0]}\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][1][0]}\"\n }\n ]\n },\n {\n \"type\": \"Column\",\n \"items\": [\n {\n \"type\": \"TextBlock\",\n \"text\": \"**#{variables('Object Variable')['tables'][0]['columns'][1]['Name']}**\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][0][1]}\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][1][1]}\"\n }\n ]\n },\n {\n \"type\": \"Column\",\n \"items\": [\n {\n \"type\": \"TextBlock\",\n \"text\": \"**#{variables('Object Variable')['tables'][0]['columns'][2]['Name']}**\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][0][2]}\"\n },\n {\n \"type\": \"TextBlock\",\n \"text\": \"#{variables('Object Variable')['tables'][0]['rows'][1][2]}\"\n }\n ]\n }\n ]\n}"
}
]
},
"runAfter": {
"Initialize_Object": [
"Succeeded"
]
},
"type": "InitializeVariable"
},
"Initialize_Object": {
"inputs": {
"variables": [
{
"name": "Object Variable",
"type": "object",
"value": {
"tables": [
{
"columns": [
{
"name": "TimeGenerated",
"type": "datetime"
},
{
"name": "_queue",
"type": "string"
},
{
"name": "_messages",
"type": "real"
}
],
"name": "PrimaryResult",
"rows": [
[
"2022-06-03T03:20:00Z",
"queue1",
8073
],
[
"2022-06-03T03:20:00Z",
"queue2",
570
]
]
}
]
}
}
]
},
"runAfter": {},
"type": "InitializeVariable"
}
},
"contentVersion": "1.0.0.0",
"outputs": {},
"parameters": {},
"triggers": {
"Recurrence": {
"evaluatedRecurrence": {
"frequency": "Month",
"interval": 12
},
"recurrence": {
"frequency": "Month",
"interval": 12
},
"type": "Recurrence"
}
}
},
"parameters": {}
}
Basically, you just need to navigate to the specific section of the JSON to retrieve what you need for each individual value.
This is an example of what I'm talking about ...
variables('Object Variable')['tables'][0]['columns'][1]['Name']
... that will get you to the name property for the second column in the first table.

Answer 2

My own solution. Essentially it's broken down into two stages: Initialize the column header and then append the rows at 2nd stage.
Pseudo code
var columns = [];
// initialize column header
for (i=0; i<input.tables[0].columns.length; i++) {
columns.add({
"type": "Column",
"items": [{
"type": "TextBlock",
"text": "**input.tables[0].columns[i].name**",
}]
});
}
// populate row
for (j=0; j<input.tables[0].rows.length; j++) {
for (i=0; i<input.tables[0].columns.length; i++) {
var thisColumn = columns[i];
thisColumn.items.add({
"type": "TextBlock",
"text": "input.tables[0].row[j][i]"
});
}
}
Full solution:
{
"definition": {
"$schema": "https://schema.management.azure.com/providers/Microsoft.Logic/schemas/2016-06-01/workflowdefinition.json#",
"actions": {
"Compose_Final": {
"inputs": {
"columns": "#variables('columns')",
"type:": "ColumnSet"
},
"runAfter": {
"Until": [
"Succeeded"
]
},
"type": "Compose"
},
"Initialize_Input": {
"inputs": {
"variables": [
{
"name": "input",
"type": "object",
"value": {
"tables": [
{
"columns": [
{
"name": "_queue",
"type": "string"
},
{
"name": "_messages",
"type": "real"
}
],
"name": "PrimaryResult",
"rows": [
[
"queue1",
8073
],
[
"queue2",
570
],
[
"queue3",
666
]
]
}
]
}
}
]
},
"runAfter": {},
"type": "InitializeVariable"
},
"Initialize_variable_columnIdx": {
"inputs": {
"variables": [
{
"name": "columnIdx",
"type": "integer",
"value": 0
}
]
},
"runAfter": {
"Initialize_variable_columns": [
"Succeeded"
]
},
"type": "InitializeVariable"
},
"Initialize_variable_columns": {
"inputs": {
"variables": [
{
"name": "columns",
"type": "array",
"value": []
}
]
},
"runAfter": {
"Initialize_Input": [
"Succeeded"
]
},
"type": "InitializeVariable"
},
"Select_Column_Header": {
"inputs": {
"from": "#variables('input')?['tables'][0]['columns']",
"select": {
"items": [
{
"text": "**#{item()['name']}**",
"type": "TextBlock"
}
],
"type": "Column"
}
},
"runAfter": {
"Initialize_variable_columnIdx": [
"Succeeded"
]
},
"type": "Select"
},
"Until": {
"actions": {
"Append_to_array_variable": {
"inputs": {
"name": "columns",
"value": {
"items": "#outputs('Compose')",
"type": "Column"
}
},
"runAfter": {
"Compose": [
"Succeeded"
]
},
"type": "AppendToArrayVariable"
},
"Compose": {
"inputs": "#union(body('Select_Column_Header')[variables('columnIdx')]['items'], body('Select'))",
"runAfter": {
"Select": [
"Succeeded"
]
},
"type": "Compose"
},
"Increment_variable": {
"inputs": {
"name": "columnIdx",
"value": 1
},
"runAfter": {
"Append_to_array_variable": [
"Succeeded"
]
},
"type": "IncrementVariable"
},
"Select": {
"inputs": {
"from": "#variables('input')?['tables'][0]['rows']",
"select": {
"text": "#item()[variables('columnIdx')]",
"type": "TextBlock"
}
},
"runAfter": {},
"type": "Select"
}
},
"expression": "#greaterOrEquals(variables('columnIdx'), length(variables('input')?['tables'][0]['columns']))",
"limit": {
"count": 60,
"timeout": "PT1H"
},
"runAfter": {
"Select_Column_Header": [
"Succeeded"
]
},
"type": "Until"
}
},
"contentVersion": "1.0.0.0",
"outputs": {},
"parameters": {},
"triggers": {
"Recurrence": {
"evaluatedRecurrence": {
"frequency": "Month",
"interval": 12
},
"recurrence": {
"frequency": "Month",
"interval": 12
},
"type": "Recurrence"
}
}
},
"parameters": {}
}