Calculating movement values from an array of balance values - mongodb

Question

Does anyone know how to calculate the difference of values between documents? So, on current document value deduct the previous document value to create a new value of the movement. Each document represents a month and year balance.
I have a set of account_balances in a document which are dated at the end of each month. They represent the general ledger from accounting app where the integration only provides the balances and not the month by month movement.
How would I calculate the difference of a balance value from an array of one document and the previous month's document?
The parameters to group together correctly are the _id.company, _id.connection, _id.object_snapshot_date and then account_balances.account_id and account_balances.value_type.
The value I want to deduct total_value of each account from 2018-05-30 from total_value from the 2018-06-31 document. There may be multiple documents in here related to an entire year.
What I want to get is the same document back but the total_value for June is the movement instead of the balance.
Thanks, Matt
Example of two documents with different months:
{
"_id" : {
"company" : " a8aa7d3f-cef8-4895-a83e-3087b4cf529c ",
"connection" : "a4b52d3a-0c00-406f-9163-4b1d52df0271",
"object_snapshot_date" : 20180603135959,
"object_schema" : "timeline-balance",
"object_class" : "trial-balance",
"object_category" : "balance",
"object_type" : "month",
"object_origin_category" : "bookkeeping",
"object_origin_type" : "accounting",
"object_origin" : "Xero"
},
"account_balances" : [
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636c96f",
"account_name" : "Sales",
"account_code" : "200",
"account_class" : "revenue",
"account_category" : "sales",
"account_group" : "",
"value_type" : "credit",
"total_value" : 29928.96,
"value_currency" : "NZD"
},
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636aa43",
"account_name" : "Cost of Goods Sold",
"account_code" : "300",
"account_class" : "expense",
"account_category" : "sales",
"account_group" : "",
"value_type" : "debit",
"total_value" : 12452.50,
"value_currency" : "NZD"
}
]
},
{
"_id" : {
"company" : " a8aa7d3f-cef8-4895-a83e-3087b4cf529c ",
"connection" : "a4b52d3a-0c00-406f-9163-4b1d52df0271",
"object_snapshot_date" : 20180503035959,
"object_schema" : "timeline-balance",
"object_class" : "trial-balance",
"object_category" : "balance",
"object_type" : "month",
"object_origin_category" : "bookkeeping",
"object_origin_type" : "accounting",
"object_origin" : "Xero"
},
"account_balances" : [
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636c96f",
"account_name" : "Sales",
"account_code" : "200",
"account_class" : "revenue",
"account_category" : "sales",
"account_group" : "",
"value_type" : "credit",
"total_value" : 24231.12,
"value_currency" : "NZD"
},
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636aa43",
"account_name" : "Cost of Goods Sold",
"account_code" : "300",
"account_class" : "expense",
"account_category" : "sales",
"account_group" : "",
"value_type" : "debit",
"total_value" : 6875.10,
"value_currency" : "NZD"
}
]
}
Expected Output would be like this:
{
"_id" : {
"company" : " a8aa7d3f-cef8-4895-a83e-3087b4cf529c ",
"connection" : "a4b52d3a-0c00-406f-9163-4b1d52df0271",
"object_snapshot_date" : 20180603135959,
"object_schema" : "timeline-balance",
"object_type" : "month",
"object_origin_category" : "bookkeeping",
"object_origin" : "Xero"
},
"account_movements" : [
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636c96f",
"account_name" : "Sales",
"account_code" : "200",
"account_class" : "revenue",
"movement" : 5697.84
},
{
"account_id" : "47cf9c6e-4ec7-4853-9efa-9e180636aa43",
"account_name" : "Cost of Goods Sales",
"account_code" : "200",
"account_class" : "revenue",
"movement" : 5577.4
}
]
}

Answer 1

I'm assuming that you can always put a filtering condition that will guarantee that only two documents remain after $match stage (like below). Then you can use $unwind to get single account_balance per document. In the next stage you can $sort by snapshot_date. Then you can $group by account_name with $push to get all balances. Since there is an assumption that there will be only two elements you can use $subtract with $arrayElemAt to get the movement.
db.col.aggregate([
{
$match: {
"_id.object_snapshot_date": {
$gte: 20180500000000,
$lte: 20180630000000
}
}
},
{
$unwind: "$account_balances"
},
{
$sort: { "_id.object_snapshot_date": 1 }
},
{
$group: {
_id: "$account_balances.account_name",
balances: { $push: "$account_balances.total_value" }
}
},
{
$project: {
_id: 0,
account_name: "$_id",
movement: { $subtract: [ { $arrayElemAt: [ "$balances", 1 ] }, { $arrayElemAt: [ "$balances", 0 ] } ] }
}
}
])
Outputs:
{ "account_name" : "Cost of Goods Sold", "movement" : 5577.4 }
{ "account_name" : "Sales", "movement" : 5697.84 }
If you need more generic solution (for more than two months) you can replace last pipeline stage with below:
{
$project: {
_id: 0,
account_name: "$_id",
movement: {
$map: {
input: { $range: [ 1, { $size: "$balances" } ] },
as: "index",
in: {
$subtract: [
{ $arrayElemAt: [ "$balances", "$$index" ] },
{ $arrayElemAt: [ "$balances", { $subtract: [ "$$index", 1 ] } ] }
]
}
}
}
}
}
This will calculate the differences for all the values in balances array using (you'll get n-1 results where n is a size of balances).