possible to use one schema's virtual fields in another virtual field (same schema)?

[FlorianML]

Hello, so wondering if this was possible or not, not having any luck. I'm running into an issue with one of the virtual fields on a schema.

We have set 2 virtuals on the schema, the first one relies on a document from a different collection and this is retrieving the document successfully. The issue is that the 2nd virtual (clientTaxYearReportCount) on the schema, uses a field that is returned from the first virtual in order to correctly populate the field. Added a snippet below to maybe better communicate the issue.

/**
 * Add virtual fields
 */
schema.virtual('clientTaxDetail', {
  foreignField: 'id',
  localField: 'clientAccountId',
  ref: 'TaxPayer',
  justOne: true,
});

schema.virtual('clientTaxYearReportCount', {
  foreignField: 'taxAccountId',
  localField: 'clientAccountId',
  ref: 'report',
  match: (doc) => ({ taxYear: doc.clientTaxDetail.taxYear, status: 'done' }),
  count: true,
});

the virtual clientTaxYearReportCount uses the clientTaxDetail.taxYear field to use to match documents from a different collection. When running, the match operation returns clientTaxDetail as undefined. Is there anyway to have the virtual fields available? I've tried adding { toJSON: true} and { toObject: true } but didn't help.

Thank you!

[vkarpov15]

It depends on how you call populate(). Remember that, if you call populate('clientTaxDetail clientTaxYearReportCount'), Mongoose will populate the two virtuals in parallel. So clientTaxDetail might not be defined yet when the match function runs.

If you wait until clientTaxDetail is populated before populating clientTaxYearReportCount as shown below, you'll get the correct result:

  const res = await Test.findOne().populate('clientTaxDetail');
  await res.populate('clientTaxYearReportCount');

Below is a complete example:

'use strict';

const mongoose = require('mongoose');
const { Schema } = mongoose;

run().catch(err => console.log(err));

async function run() {
  await mongoose.connect('mongodb://localhost:27017/test');
  
  const schema = new Schema({
    clientAccountId: String
  });
  
  schema.virtual('clientTaxDetail', {
    foreignField: 'id',
    localField: 'clientAccountId',
    ref: 'TaxPayer',
    justOne: true,
  });
  
  schema.virtual('clientTaxYearReportCount', {
    foreignField: 'taxAccountId',
    localField: 'clientAccountId',
    ref: 'Report',
    match: (doc) => ({ taxYear: doc.clientTaxDetail.taxYear, status: 'done' }),
    count: true, 
  });
  
  const Test = mongoose.model('Test', schema);
  const TaxPayer = mongoose.model('TaxPayer', new Schema({
    id: String,
    taxYear: String,
    clientAccountId: String
  }));
  const Report = mongoose.model('Report', new Schema({ taxAccountId: String, status: String, taxYear: String }));

  await Report.create({ taxAccountId: 'testid1', status: 'done', taxYear: '2021' });
  await TaxPayer.create({ id: 'testid2', taxYear: '2021', clientAccountId: 'testid1' });
  await Test.create({ clientAccountId: 'testid2' });

  const res = await Test.findOne().populate('clientTaxDetail');
  await res.populate('clientTaxYearReportCount');
  console.log(res.toObject({ virtuals: true }));
}

Another alternative would be to make clientTaxYearReportCount a populated virtual on the TaxPayer document, and using deep populate.