/**
 * Tests $unionWith with various levels of nesting and stages.
 */

(function() {
"use strict";
load("jstests/aggregation/extras/utils.js");  // arrayEq
load("jstests/libs/fixture_helpers.js");      // For FixtureHelpers.

const testDB = db.getSiblingDB(jsTestName());
const collA = testDB.A;
collA.drop();
const collB = testDB.B;
collB.drop();
const collC = testDB.C;
collC.drop();
const collD = testDB.D;
collD.drop();
const docsPerCollection = 5;
for (let i = 0; i < docsPerCollection; i++) {
    assert.commandWorked(collA.insert({a: i, val: i, groupKey: i}));
    assert.commandWorked(collB.insert({b: i, val: i * 2, groupKey: i}));
    assert.commandWorked(collC.insert({c: i, val: 10 - i, groupKey: i}));
    assert.commandWorked(collD.insert({d: i, val: i * 3, groupKey: i}));
}

function checkResults(resObj, expectedResult) {
    assert.commandWorked(resObj);
    assert(anyEq(resObj.cursor.firstBatch, expectedResult),
           "Expected:\n" + tojson(expectedResult) + "Got:\n" + tojson(resObj.cursor.firstBatch));
}

function getDocsFromCollection(collObj, proj = null) {
    if (proj == null) {
        return collObj.find().toArray();
    } else {
        return collObj.aggregate({"$addFields": proj}).toArray();
    }
}
// Test a two collection union.
let resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collB));
checkResults(
    testDB.runCommand(
        {aggregate: collA.getName(), pipeline: [{$unionWith: collB.getName()}], cursor: {}}),
    resSet);
// Test a sequential four collection union.
resSet = getDocsFromCollection(collA).concat(
    getDocsFromCollection(collB), getDocsFromCollection(collC), getDocsFromCollection(collD));
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [
        {$unionWith: collB.getName()},
        {$unionWith: collC.getName()},
        {$unionWith: collD.getName()}
    ],
    cursor: {}
}),
             resSet);
// Test a nested four collection union.
// resSet should be the same.
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{
        $unionWith: {
            coll: collB.getName(),
            pipeline:
                [{$unionWith: {coll: collC.getName(), pipeline: [{$unionWith: collD.getName()}]}}]
        }
    }],
    cursor: {}
}),
             resSet);
// Test that a sub-pipeline is applied to the correct documents.
resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collB, {x: 3}));
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{$unionWith: {coll: collB.getName(), pipeline: [{"$addFields": {x: 3}}]}}],
    cursor: {}
}),
             resSet);
// Test that for multiple nested unions sub-pipelines are applied to the correct documents.
resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collB, {x: 3}),
                                             getDocsFromCollection(collC, {x: 3, y: 4}),
                                             getDocsFromCollection(collD, {x: 3, y: 4, z: 5}));
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{
        $unionWith: {
            coll: collB.getName(),
            pipeline: [
                {
                    $unionWith: {
                        coll: collC.getName(),
                        pipeline: [
                            {
                                $unionWith:
                                    {coll: collD.getName(), pipeline: [{"$addFields": {z: 5}}]}
                            },
                            {"$addFields": {y: 4}}
                        ]
                    }
                },
                {"$addFields": {x: 3}}
            ]
        }
    }],
    cursor: {}
}),
             resSet);
resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collB, {x: 3}),
                                             getDocsFromCollection(collC, {x: 3, y: 4}),
                                             getDocsFromCollection(collD, {x: 3, z: 5}));
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{
        $unionWith: {
            coll: collB.getName(),
            pipeline: [
                {$unionWith: {coll: collC.getName(), pipeline: [{"$addFields": {y: 4}}]}},
                {$unionWith: {coll: collD.getName(), pipeline: [{"$addFields": {z: 5}}]}},
                {"$addFields": {x: 3}}
            ]
        }
    }],
    cursor: {}
}),
             resSet);
// Test with $group. Set MaxMemory low to force spillToDisk.
FixtureHelpers.runCommandOnEachPrimary({
    db: testDB.getSiblingDB("admin"),
    cmdObj: {
        setParameter: 1,
        "internalDocumentSourceGroupMaxMemoryBytes": 1,
    }
});
resSet =
    [{_id: 0, sum: 0}, {_id: 1, sum: 3}, {_id: 2, sum: 6}, {_id: 3, sum: 9}, {_id: 4, sum: 12}];
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{$unionWith: collB.getName()}, {"$group": {_id: "$groupKey", sum: {$sum: "$val"}}}],
    cursor: {},
    // we allow disk use to test the case where the pipeline must run on a shard in the sharded
    // passthrough.
    allowDiskUse: true
}),
             resSet);
// Test a $group that sums in both the inner and outer pipeline.
resSet =
    [{_id: 0, sum: 0}, {_id: 1, sum: 21}, {_id: 2, sum: 2}, {_id: 3, sum: 3}, {_id: 4, sum: 4}];
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [
        {
            $unionWith: {
                coll: collB.getName(),
                pipeline: [
                    {"$group": {_id: "$groupKey", val: {$sum: "$val"}}},
                    {"$addFields": {groupKey: 1}}
                ]
            }
        },
        {"$group": {_id: "$groupKey", sum: {$sum: "$val"}}}
    ],
    allowDiskUse: true,
    cursor: {}
}),
             resSet);
resSet = getDocsFromCollection(collA).concat(
    [{_id: 0, sum: 0}, {_id: 1, sum: 2}, {_id: 2, sum: 4}, {_id: 3, sum: 6}, {_id: 4, sum: 8}]);
// Test a $group in the inner pipeline.
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [
        {
            $unionWith: {
                coll: collB.getName(),
                pipeline: [
                    {"$group": {_id: "$groupKey", sum: {$sum: "$val"}}},
                ]
            }
        },
    ],
    allowDiskUse: true,
    cursor: {}
}),
             resSet);

// Test that a $group within a $unionWith sub-pipeline correctly fails if it needs to spill but
// 'allowDiskUse' is false.
assert.commandFailedWithCode(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [
        {
            $unionWith: {
                coll: collB.getName(),
                pipeline: [
                    {"$group": {_id: "$groupKey", val: {$sum: "$val"}}},
                    {"$addFields": {groupKey: 1}}
                ]
            }
        },
    ],
    allowDiskUse: false,
    cursor: {}
}),
                             ErrorCodes.QueryExceededMemoryLimitNoDiskUseAllowed);

// Reset to default value.
FixtureHelpers.runCommandOnEachPrimary({
    db: testDB.getSiblingDB("admin"),
    cmdObj: {
        setParameter: 1,
        "internalDocumentSourceGroupMaxMemoryBytes": 100 * 1024 * 1024,
    }
});

// Test with $limit and sort in a sub-pipeline.
const setBResult = collB.find().sort({b: 1}).limit(2).toArray();
resSet = getDocsFromCollection(collA).concat(setBResult);
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{$unionWith: {coll: collB.getName(), pipeline: [{"$sort": {b: 1}}, {"$limit": 2}]}}],
    cursor: {}
}),
             resSet);

// Test that we get the correct number of documents when using a getmore.
resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collD)).length;
let fullResults = collA
                      .aggregate([{$unionWith: collD.getName()}],
                                 {cursor: {batchSize: Math.floor(docsPerCollection / 2)}})
                      .itcount();
assert.eq(resSet,
          fullResults,
          "Expected: " + resSet + " Got: " + fullResults.length + " from " + tojson(fullResults));

// Test with a sub-pipeline and a getMore before the initial collection is exhausted.
resSet = getDocsFromCollection(collA).concat(getDocsFromCollection(collD, {x: 3})).length;
fullResults =
    collA
        .aggregate([{$unionWith: {coll: collD.getName(), pipeline: [{"$addFields": {x: 3}}]}}],
                   {cursor: {batchSize: Math.floor(docsPerCollection / 2)}})
        .itcount();
assert.eq(resSet,
          fullResults,
          "Expected: " + resSet + " Got: " + fullResults.length + " from " + tojson(fullResults));

// Test when a getMore occurs after the initial collection is exhausted. resSet remains the same as
// only batchSize has changed.
fullResults =
    collA
        .aggregate([{$unionWith: {coll: collD.getName(), pipeline: [{"$addFields": {x: 3}}]}}],
                   {cursor: {batchSize: docsPerCollection + 1}})
        .itcount();
assert.eq(resSet,
          fullResults,
          "Expected: " + resSet + " Got: " + fullResults.length + " from " + tojson(fullResults));

// Test that $unionWith on a non-existent collection succeeds.
resSet = getDocsFromCollection(collA);
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline: [{$unionWith: "nonExistentCollectionAlpha"}],
    cursor: {}
}),
             resSet);
checkResults(testDB.runCommand({
    aggregate: collA.getName(),
    pipeline:
        [{$unionWith: {coll: "nonExistentCollectionAlpha", pipeline: [{"$addFields": {ted: 5}}]}}],
    cursor: {}
}),
             resSet);
checkResults(testDB.runCommand({
    aggregate: "nonExistentCollectionAlpha",
    pipeline: [{$unionWith: collA.getName()}],
    cursor: {}
}),
             resSet);
})();