mongodb aggregation select specific document in group - mongodb

Question

I need a bit help with mongodb aggregation.
first I have a $match to get filter some specific documents.
then I group by a field I need them grouped in.
the group I need to select a document where field value is ... and get that document as main data.
{"$match": {"$and": [
{chain: chain},
{dex: dex}
]}};
{$group: {
_id: "$pairAddress",
allChange: {"$push": "$$ROOT"},
baseToken: {$last: '$baseToken'},
txCount: {the document with timeframe inside this group 86400.txCount}
}},
{$sort: {txCount: -1}}
{$skip: 0}
{$limit: 100}
the group consist of documents with different timeframes, I need to somehow select a specific timeframe and add fields to the group from that timeframe. for example each timeframe has a different amount of txCount after group I want to sort by txCount and limit the amount and use skip for some pagination.
the problem is in selecting a document from that group with the specific timeframe.
anyone who could help me a bit to the right direction that would be awesome.
Here an example of how data is stored in the database and what I would like the result to be.
const document = {
_id: '3567356735672467',
pairAddress: '0x45jk6v34jy5634jkh5v6kj4h5v62j4h56', // group by pair address
baseToken: '0x456jn345k6hb4k5h6b3khb65k3hb56k3h4b6',
resolution: 86400, // a pair address has 6 documents with each a own timeframe 300, 900, 1800, 3600, 43200, 86400
base0: true,
txCount: 26,
buyCount: 10,
sellCount:16,
buyVolume: '2342354.345',
sellVolume: '1234.34',
volume: '1232352.345',
change: '12.34',
positive: true,
time: 1676865981,
chain: 'ETH',
dex: 'SUS',
price: '12.45',
};
const result = [
{
_id: "0x45jk6v34jy5634jkh5v6kj4h5v62j4h56",
allChange: {"$push": "$$ROOT"}, // array of all documents/timeframes for a pairAddress
selectedTxAmount: 26, // this needs to be the document with selected timeframe example 86400, selected from the group is must match the pairAddress
}
];
Maybe its possible to change the aggregation to make it work and faster.
match all timeframes, dex and chain.
sort by txCount.
skip X amount.
limit to 100
and return all document with a field containing all timestamps per the pairAddress left after the aggregation.
Currently thanks to #1sina1 I got this and it works.
{"$match": {"$and": [
{"chain": chain},
{"dex": dex}
]}},
{$group: {
_id: "$pairAddress",
allChange: {"$push": "$$ROOT"},
baseToken: {$last: '$baseToken'},
txCount: {
"$push": {
"$cond": {
"if": {
"$eq": [
"$resolution",
43200
]
},
"then": "$txCount",
"else": "$$REMOVE"
}
}
}
}},
{$sort: {txCount: -1}},
{$skip: parseInt(page) * 100},
{$limit: 100},
But I think there might be a way to do it just a bit different now we first group all (which is about 20k documents) I am only interested in 100, so maybe first match to timeframe/resolution then sort, skip, limit and then I just need from those 100 pairAddress all the according timeframes/resolutions for each as a flied allChange.