選取

說明

產生新文件，方法是參照現有欄位的子集，或是將欄位指派給指定運算式的結果。

範例

Web

const result = await execute(db.pipeline()
  .collection("books")
  .select(field("soldBooks").multiply(field("price")).round().as("partialRevenue"))
  .aggregate(field("partialRevenue").sum().as("totalRevenue"))
  );
Swift
let result = try await db.pipeline()
  .collection("books")
  .select([Field("soldBooks").multiply(Field("price")).round().as("partialRevenue")])
  .aggregate([Field("partialRevenue").sum().as("totalRevenue")])
  .execute()

Kotlin

val result = db.pipeline()
    .collection("books")
    .select(Expression.multiply(field("soldBooks"), field("price")).round().alias("partialRevenue"))
    .aggregate(AggregateFunction.sum("partialRevenue").alias("totalRevenue"))
    .execute()

Java

Task<Pipeline.Snapshot> result = db.pipeline()
    .collection("books")
    .select(Expression.multiply(field("soldBooks"), field("price")).round().alias("partialRevenue"))
    .aggregate(AggregateFunction.sum("partialRevenue").alias("totalRevenue"))
    .execute();
Python
from google.cloud.firestore_v1.pipeline_expressions import Field

result = (
    client.pipeline()
    .collection("books")
    .select(
        Field.of("soldBooks")
        .multiply(Field.of("price"))
        .round()
        .as_("partialRevenue")
    )
    .aggregate(Field.of("partialRevenue").sum().as_("totalRevenue"))
    .execute()
)
Java
Pipeline.Snapshot result =
    firestore
        .pipeline()
        .collection("books")
        .select(round(multiply(field("soldBooks"), field("price"))).as("partialRevenue"))
        .aggregate(sum("partialRevenue").as("totalRevenue"))
        .execute()
        .get();

行為

「選取階段」的位置

選取階段的使用時間沒有限制，但如果欄位未納入選取階段，管道的後續階段就無法存取這些欄位。舉例來說，如要從下列資料集選取加拿大所有城市的 name 和 location 欄位：

Node.js

await db.collection("cities").doc("SF").set({
  name: "San Francisco",
  population: 800000,
  location: {country: "USA", state: "California"}
});

await db.collection("cities").doc("TO").set({
  name: "Toronto",
  population: 3000000,
  location: {country: "Canada", province: "Ontario"}
});

您可以使用下列管道：

Node.js

const names = await db.pipeline()
  .collection("/cities")
  .where(equal(field("location.country"), "Canada"))
  .select(stringConcat(field("name"), ", ", field("location.country")).as("name"), "population")
  .execute();

這會產生下列文件：

{ name: "Toronto, Canada", population: 3000000 },

不過，如果 select(...) 階段是放在 where(...) 階段之前，例如：

Node.js

const names = await db.pipeline()
  .collection("/cities")
  .select(stringConcat(field("name"), ",", field("location.country")).as("name"), "population")
  .where(equal(field("location.country"), "Canada"))
  .execute();

由於 location.country 已在 where(...) 階段執行前從文件中移除，因此不會產生任何文件。

選取巢狀欄位

select(...) 階段可用於從地圖和陣列中選取巢狀欄位。舉例來說，如要從下列文件中選取巢狀 country 欄位和 landmarks 陣列的第一個項目：

Node.js

await db.collection("cities").doc("SF").set({
  name: "San Francisco",
  population: 800000,
  location: { country: "USA", state: "California" },
  landmarks: [ "Golden Gate Bridge", "Alcatraz" ]
});

await db.collection("cities").doc("TO").set({
  name: "Toronto",
  population:  3000000,
  province: "ON",
  location: { country: "Canada", province: "Ontario" },
  landmarks: [ "CN Tower", "Casa Loma" ]
});

await db.collection("cities").doc("AT").set({
  name: "Atlantis",
  population: null
});

您可以使用下列管道：

Node.js

const locations = await db.pipeline()
  .collection("/cities")
  .select(
    field("name").as("city"),
    field("location.country").as("country"),
    field("landmarks").offset(0).as("topLandmark"))
  .execute();

這會產生下列文件：

{ city: "San Francisco", country: "USA", topLandmark: "Golden Gate Bridge" },
{ city: "Toronto", country: "Canada", topLandmark: "CN Tower" },
{ city: "Atlantis" }

如果沒有巢狀對應值或陣列值，就不會納入產生的文件中。選取階段中的陣列和對應存取行為，分別與 offset(...) 和 get_field(...) 函式相同。

指派巢狀欄位

運算式的結果也可以指派給巢狀欄位，因此 select(...) 階段可以從前一個階段傳回巢狀欄位的子集。舉例來說，下列項目可用於確保只傳回 city 和 state 資訊，同時保留文件的原始形狀：

Node.js

const results = await db.pipeline()
  .collection("/users")
  .addFields(
    field("__name__"),
    field("address.city").as("address.city"),
    field("address.state").as("address.state"))
  .execute();