我使用" int-file:inbound-channel-adapter"加载存在于目录中的文件。我喜欢按顺序处理文件:这意味着当第一个文件的处理完成后,我加载第二个文件......等等。
我看到sample,但我无法预测处理一个文件所需的时间,这取决于其大小。
我的源代码:
<int-file:inbound-channel-adapter
directory="${directory.files.local}" id="filesIn" channel="channel.filesIn">
<int:poller fixed-delay="1000" max-messages-per-poll="1" />
</int-file:inbound-channel-adapter>
文件的过程是文件:inbound-channel-adapter ---&gt; transformer ---&gt; splitter ----&gt; http:outbound-gateway ---&gt; outbound-mail-adapter- ---&gt;文件处理完成,所以此时,我要处理下一个文件。
我的项目配置太复杂了。 下面,我向您展示更多配置: 配置的第一部分是:
<int-sftp:inbound-channel-adapter id="sftpInbondAdapter"
auto-startup="true" channel="receiveChannel" session-factory="sftpSessionFactory"
local-directory="file:${directory.files.local}" remote-directory="${directory.files.remote}"
auto-create-local-directory="true" delete-remote-files="true"
filename-regex=".*\.txt$">
<int:poller fixed-delay="${sftp.interval.request}"
max-messages-per-poll="-1" />
</int-sftp:inbound-channel-adapter>
<!-- <int:poller cron="0 * 17 * * ?"></int:poller> -->
<int-file:inbound-channel-adapter
filter="compositeFileFilter" directory="${directory.files.local}" id="filesIn"
channel="channel.filesIn" prevent-duplicates="true">
<int:poller fixed-delay="1000" max-messages-per-poll="1" />
</int-file:inbound-channel-adapter>
<int:transformer input-channel="channel.filesIn"
output-channel="channel.file.router" ref="fileTransformer" method="transform" />
<int:recipient-list-router id="fileRouter"
input-channel="channel.file.router">
<int:recipient channel="channel.empty.files"
selector-expression="payload.length()==0" />
<int:recipient channel="channel.filesRejected"
selector-expression="payload.toString().contains('rejected')" />
<int:recipient channel="toSplitter"
selector-expression="(payload.length()>0) and(!payload.toString().contains('rejected'))" />
</int:recipient-list-router>
然后从通道tosplitter,我的程序逐行读取文件:
<int-file:splitter input-channel="toSplitter"
output-channel="router" requires-reply="false" />
<int:recipient-list-router id="recipentRouter"
input-channel="router">
<int:recipient channel="channelA"
selector-expression="headers['file_name'].startsWith('${filenameA.prefix}')" />
<int:recipient channel="channelB"
selector-expression="headers['file_name'].startsWith('${filenameB.prefix}')" />
</int:recipient-list-router>
每个通道A和B应为每一行调用两个不同的WS。 每个文件对文件A使用asynch调用ws代码:
<int:header-enricher input-channel="channelA"
output-channel="channelA.withHeader">
<int:header name="content-type" value="application/json" />
<int:header name="key1" expression="payload.split('${line.column.separator}')[0]" />
<int:header name="key2" expression="payload"></int:header>
</int:header-enricher>
<int:transformer input-channel="channelA.withHeader"
output-channel="channelA.request" ref="imsiMsgTransformer"
method="transform">
</int:transformer>
<int:channel id="channelA.request">
<int:queue capacity="10" />
<int-http:outbound-gateway id="maspUpdatorSimChangedGateway"
request-channel="channelA.request"
url="${url}"
http-method="PUT" expected-response-type="java.lang.String" charset="UTF-8"
reply-timeout="${ws.reply.timeout}" reply-channel="channelA.reply">
<int-http:uri-variable name="foo" expression="headers['key1']" />
<int:poller fixed-delay="1000" error-channel="channelA.error"
task-executor="executorA" />
<int-http:request-handler-advice-chain>
<int:retry-advice max-attempts="${ws.max.attempts}"
recovery-channel="recovery.channelA">
<int:fixed-back-off interval="${ws.interval.attempts}" />
</int:retry-advice>
</int-http:request-handler-advice-chain>
</int-http:outbound-gateway>
<int:service-activator input-channel="recovery.channelA"
ref="updateImsiHttpResponseErrorHandler" method="handleMessage" output-channel="updateImsi.channel.error.toenricher">
</int:service-activator>
<int:service-activator input-channel="channelA.reply"
ref="updateImsiHttpResponseMessageHandler" method="handleMessage">
<int:poller fixed-delay="1000"></int:poller>
</int:service-activator>
在(回复频道和恢复频道)的每个激活器中,我计算文件的进度,直到文件完成此时我应该加载第二个文件A2或文件B ......等等。
答案 0 :(得分:1)
这是默认行为,只要
task-executor
(你的没有)。DirectChannel
s(默认值) - 这表示没有QueueChannel
或ExecutorChannel
(即没有task-executor
或<queue/>
在频道上。)在这种情况下,直到当前轮询结束时才会考虑下一个轮询 - 流程在轮询器线程上运行,并且一次只能进行一次轮询。
在完全处理当前文件之前,fixed-delay
才会启动。
修改强>
如果您需要对流使用异步处理,则需要使用Conditional Poller或简单PollSkipAdvice。
您将提供一个PollSkipStrategy
实现,在文件完成之前将返回false。
这样,随后的民意调查将被跳过,直到你决定。
<强> EDIT2 强>
像这样......
/*
* Copyright 2015 the original author or authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.springframework.integration.scheduling;
/**
* @author Gary Russell
* @since 4.3
*
*/
public class SimplePollSkipStrategy implements PollSkipStrategy {
private volatile boolean skip;
@Override
public boolean skipPoll() {
return this.skip;
}
public void skipPolls() {
this.skip = true;
}
public void reset() {
this.skip = false;
}
}
<bean/>
到您的上下文。PollSkipAdvice
skipPolls()
。reset()
。