Trying to aggregate based on substring matches in mongodb 3.2 - mongodb

Question

Let's say my collection has documents with ExpName field and Rname field. Expname are all of the type - exp_1, exp_2 etc. Rname is a character string with 4 dashes for example. "As-34rt-d3r5-4453f-er4"
I need to aggregate based on experiment name and removing the text between the last two dashes. In the example I gave above that would be "As-34rt-d3r5"
question 1) how do i incorporate this in one table?
question 2) i solved this in a dirty fashion for one exp, because it seemed like the number of characters was almost the same, so I could just take the first 13 characters which seemed like it was the the substring omitting the last two dashes. Is there a correct way to do this if the text was not so uniform?
db.getCollection('rest01').aggregate(
{$match : {ExpName : "exp_1"}},
{$group: {_id :"$ExpName",_id : {$substr : ["$RName", 0,13]}, total: { $sum:1 }}
})
Ideally I would like to have a result that says Expname, Rnamesubstring, count. This code snippet was for exp_1 one alone. Is it even possible to get it all in one result?

Answer 1

Here is how you could do that:
db.getCollection('rest01').aggregate({
$project: {
"ExpName": 1,
"splitRName": { $split: [ "$RName", "-" ] } // add an array with the constituents of your dash-delimited string id as a new field "splitRName"
}
}, {
$group: {
_id: { // our group id shall be made up of both...
"ExpName": "$ExpName", // ...the "ExpName" field...
"Rnamesubstring": { // and some parts of the "RName" field
$concat:
[
{ $arrayElemAt: [ "$splitRName", 0 ] },
"-",
{ $arrayElemAt: [ "$splitRName", 1 ] },
"-",
{ $arrayElemAt: [ "$splitRName", 2 ] }
]
}
},
total: { $sum: 1 }
}
})
In case you want to do it in MongoDB v3.2 (as stated in your comment), here is something that is not exactly pretty but works:
db.getCollection('rest01').aggregate({
$group: {
_id: { // our group id shall be made up of both...
"ExpName": "$ExpName", // ...the "ExpName" field...
"Rnamesubstring": {
$substr:
[
"$RName",
0,
{
$ifNull:
[
{
$arrayElemAt:
[{
$filter: {
input: {
$map: {
input: [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 /* add numbers as required */ ],
as: "index",
in: {
$cond: {
if: { $eq: [ "-", { $substr: [ "$RName", "$$index", 1 ] } ] }, // if the string we look at is a dash...
then: "$$index", // ...then let's remember it
else: null // ...otherwise ignore it
}
}
}
},
as: "item",
cond: { $ne: [ null, "$$item" ] } // get rid of all null values
}
},
2 ] // we want the position of the third dash in the string (only)
},
1000 // in case of a malformed RName (wrong number of dashes or completely missing) we want the entire substring
]
}
]
}
},
total: { $sum: 1 }
}
})
Update 2: You seem to be having some data related issues as per your comments (so either missing RName values or improperly structured ones, i.e. without the required number of sections with dashes in between). I have updated the above statement for v3.2 to deal with these rows. You may want to find out, though, which rows actually cause this behaviour. They can be easily identified using the following statement:
db.getCollection('rest01').aggregate({
$project: {
_id: 1,
RName: 1,
"Rnamesubstring": {
$arrayElemAt:
[{
$filter: {
input: {
$map: {
input: [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 /* add numbers as required */ ],
as: "index",
in: {
$cond: {
if: { $eq: [ "-", { $substr: [ "$RName", "$$index", 1 ] } ] }, // if the string we look at is a dash...
then: "$$index", // ...then let's remember it
else: null // ...otherwise ignore it
}
}
}
},
as: "item",
cond: { $ne: [ null, "$$item" ] } // get rid of all null values
}
},
2 ] // we want the position of the third dash in the string (only)
}
}
}, {
$match: { "Rnamesubstring": { $exists:false } }
})