使用Apache Pulsar：从指定的消息ID到结束消息ID读取/使用消息

在IT行业这个发展更新速度很快的行业，只有不停止的学习，才不会被行业所淘汰。如果你是Golang学习者，那么本文《使用Apache Pulsar：从指定的消息ID到结束消息ID读取/使用消息》就很适合你！本篇内容主要包括##content_title##，希望对大家的知识积累有所帮助，助力实战开发！

使用 kafka，我可以指定一个整数消息 id 来开始消费，并指定一个结束消息来停止，例如如下：

kafkacat -b kafka:9092 -t messages -o 11000 -c 11333

但是，指定整数开始和停止消息的相同功能在 apache pulsar 中似乎不可用！

公平地说，如果已经跟踪并以字节格式保存了开始消息 id 和结束消息 id，则可以使用非常复杂的过程来指定开始消息 id 和结束消息 id，这必然会影响性能和代码复杂性。

如本例所示：

client, err := NewClient(pulsar.ClientOptions{
    URL: lookupURL,
})

if err != nil {
    log.Fatal(err)
}
defer client.Close()

topic := "topic-1"
ctx := context.Background()

// create producer
producer, err := client.CreateProducer(pulsar.ProducerOptions{
    Topic:           topic,
    DisableBatching: true,
})
if err != nil {
    log.Fatal(err)
}
defer producer.Close()

// send 10 messages
msgIDs := [10]MessageID{}
for i := 0; i < 10; i++ {
    msgID, err := producer.Send(ctx, &pulsar.ProducerMessage{
        Payload: []byte(fmt.Sprintf("hello-%d", i)),
    })
    assert.NoError(t, err)
    assert.NotNil(t, msgID)
    msgIDs[i] = msgID
}

// create reader on 5th message (not included)
reader, err := client.CreateReader(pulsar.ReaderOptions{
    Topic:          topic,
    StartMessageID: msgIDs[4],
})

if err != nil {
    log.Fatal(err)
}
defer reader.Close()

// receive the remaining 5 messages
for i := 5; i < 10; i++ {
    msg, err := reader.Next(context.Background())
    if err != nil {
    log.Fatal(err)
}

// create reader on 5th message (included)
readerInclusive, err := client.CreateReader(pulsar.ReaderOptions{
    Topic:                   topic,
    StartMessageID:          msgIDs[4],
    StartMessageIDInclusive: true,
})

if err != nil {
    log.Fatal(err)
}
defer readerInclusive.Close()

但是，对于多个并发读取器来说，这很复杂且不可靠（或复杂），并且需要使用外部构造来跟踪已处理的消息，然后才能使用开始/结束语义检索消息。

有什么方法可以实现这一点（最好通过golang）

正确答案

我发现以下简单方法就足够了（概念验证脚本）：

package main

import (
    "context"
    "fmt"
    "io/ioutil"
    "log"
    "os"
    "strconv"
    "strings"
    "time"

    "github.com/apache/pulsar-client-go/pulsar"
)

func writeBytesToFile(f string, byteSlice []byte) int {
    // Open a new file for writing only

    f = "./data/" + f

    file, err := os.OpenFile(
        f,
        os.O_WRONLY|os.O_TRUNC|os.O_CREATE,
        0666,
    )
    if err != nil {
        log.Fatal(err)
    }
    defer file.Close()

    // Write bytes to file
    bytesWritten, err := file.Write(byteSlice)

    if err != nil {
        log.Fatal(err)
    }

    log.Printf("Wrote %d bytes.\n", bytesWritten)

    return bytesWritten
}

func readBackByEntryId(msgDir string, msgIndex string) (yourBytes []byte) {

    //We know the file name by convention
    fname := msgDir + "/" + msgIndex + ".dat"

    yourBytes, err := ioutil.ReadFile(fname)

    if err != nil {
        log.Printf("error reading %s", fname)
        return nil
    }

    return yourBytes
}

func getFiles(aDir string) []string {

    var theFiles []string

    files, err := ioutil.ReadDir("./data/")

    if err != nil {
        log.Fatal(err)
    }

    for _, f := range files {

        theFiles = append(theFiles, f.Name())

    }

    return theFiles
}

func streamAll(reader pulsar.Reader, startMsgIndex int64, stopMsgIndex int64) {

    read := false

    for reader.HasNext() {

        msg, err := reader.Next(context.Background())

        if err != nil {
            log.Fatal(err)
        }

        //can I access the details of the message ? yes
        fmt.Printf("%v -> %#v\n", msg.ID().EntryID(), msg.ID())

        //Can i serialize into bytes? Yes
        myBytes := msg.ID().Serialize()

        //Can I store it somewhere? Perhaps a map ? or even on disk in a file ?
        //In other words: Can I write a byte[] slice to a file? Yes!
        msgIndex := msg.ID().EntryID()

        if msgIndex == startMsgIndex {
            fmt.Println("start read: ", msgIndex)
            read = true
        }

        if msgIndex > stopMsgIndex {
            fmt.Println("stop reading: ", msgIndex)
            read = false
        }

        if read == false {

            fmt.Println("skipping ", msgIndex)

        } else {

            fname := strconv.FormatInt(msgIndex, 10) + ".dat"

            fmt.Println("written bytes: ", writeBytesToFile(fname, myBytes))

            fmt.Printf("Received message msgId: %#v -- content: '%s' published at %v\n",
                msg.ID(), string(msg.Payload()), msg.PublishTime())

        }

        /*
            //FYI - to save and reread a msgId from store: https://githubmemory.com/@storm-5
            msgId := msg.ID()
            msgIdBytes := msgId.Serialize()
            idNew, _ := pulsar.DeserializeMessageID(msgIdBytes)

            readerInclusive, err := client.CreateReader(pulsar.ReaderOptions{
                Topic:                   "ragnarok/transactions/requests",
                StartMessageID:          idNew,
                StartMessageIDInclusive: true,
            })
        */
    }

}

func retrieveRange(client pulsar.Client) {

    someFiles := getFiles("./data/")

    for _, f := range someFiles {

        fIndex := strings.Split(f, ".")[0]

        fmt.Println("re-reading message index -> ", fIndex)

        msgIdBytes := readBackByEntryId("./data", fIndex)

        fmt.Printf("boom -> %#v\n", msgIdBytes)

        idNew, err := pulsar.DeserializeMessageID(msgIdBytes)

        if err != nil {
            log.Fatal(err)
        }

        fmt.Println("Got message entry id => ", idNew.EntryID())

        readerInclusive, err := client.CreateReader(pulsar.ReaderOptions{
            Topic:                   "ragnarok/transactions/requests",
            StartMessageID:          idNew,
            StartMessageIDInclusive: true,
        })

        if err != nil {
            log.Fatal(err)
        }

        defer readerInclusive.Close()

        //defer readerInclusive.Close()
        fmt.Println("bleep!")

        msg, err := readerInclusive.Next(context.Background())

        if err != nil {
            log.Fatal(err)
        }

        //fmt.Println("retrieved message -> ", string(msg.Payload()))
        fmt.Printf("Retrieved message ID message msgId: %#v -- content: '%s' published at %v\n",
            msg.ID(), string(msg.Payload()), msg.PublishTime())

    }
}

func main() {

    client, err := pulsar.NewClient(
        pulsar.ClientOptions{
            URL:               "pulsar://localhost:6650",
            OperationTimeout:  30 * time.Second,
            ConnectionTimeout: 30 * time.Second,
        })

    if err != nil {
        log.Fatalf("Could not instantiate Pulsar client: %v", err)
    }

    defer client.Close()

    reader, err := client.CreateReader(pulsar.ReaderOptions{
        Topic:          "ragnarok/transactions/requests",
        StartMessageID: pulsar.EarliestMessageID(),
    })

    if err != nil {
        log.Fatal(err)
    }

    defer reader.Close()

    if err != nil {
        log.Fatal(err)
    }

    var startMsgId int64 = 55
    var stopMsgId int64 = 66

    //stream all the messages from the earliest to latest
    //pick a subset between a start and stop id
    streamAll(reader, startMsgId, stopMsgId)

    //retrieve the picked range
    retrieveRange(client)

}

好了，本文到此结束，带大家了解了《使用Apache Pulsar：从指定的消息ID到结束消息ID读取/使用消息》，希望本文对你有所帮助！关注golang学习网公众号，给大家分享更多Golang知识！