好記憶不如爛筆頭，能記下點東西，就記下點，有時間拿出來看看，也會發覺不一樣的感受.
簡單描述一下自己通過學習nifi和查看源碼，然後自己通過nifi的體現很基礎的實現自己的業務的處理器開發。
適合剛入門的看，有nifi基礎的，可以跳過！
目錄
1.Json文檔的輸入端，需要一個GenerateFlowFile處理器,做如下配置:
一、 獲取一個json文件的文本信息.
假定我們需要在一段json文件中，獲取json串中的某個key所對應的value的值，那麼在nifi的架子上，該如何操作吶？共分爲三步：
1.Json文檔的輸入端，需要一個GenerateFlowFile處理器,做如下配置:
Custom text 存放的是目標json串.json如下所示:{"error_code":0,"reason":"ok","result":{"items":[{"regStatus":"存續","estiblishTime":1495555200000,"regCapital":"","pencertileScore":4902,"type":1,"legalPersonName":"溫旭穎","toco":2,"legalPersonId":2051255554,"name":"陝西蜂窩科技股份有限公司","logo":"","alias":"蜂窩科技","id":3053414776,"category":"723","personType":1,"base":"han"},{"regStatus":"註銷","estiblishTime":1473264000000,"regCapital":"","pencertileScore":3860,"type":1,"legalPersonName":"常青","toco":8,"legalPersonId":1911055314,"name":"陝西蜂窩科技股份有限公司","logo":"","alias":"蜂窩科技","id":2958332903,"category":"721","personType":1,"base":"xj"}],"total":18}}
2.拖拽自己實現的處理器
根據要求配置好自己的json_path的路徑，這樣方便獲取數據.如: $.result.items[*]
自己處理器的主要代碼是：
@Tags({"first-example:fetch value from json string"})
@SideEffectFree
@CapabilityDescription("fetch value from json string.")
public class FirstProcessor extends AbstractProcessor {
private List<PropertyDescriptor> properties;
private Set<Relationship> relationships;
private final String arrayFlag="true";
/**
* json路徑.
*/
public static final PropertyDescriptor JSON_PATH = new PropertyDescriptor.Builder()
.name("Json Path")
.required(true)
.description("json path value,such as:$.test")
.addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
.build();
/**
* json路徑.
*/
public static final PropertyDescriptor ARRAY_FLAG = new PropertyDescriptor.Builder()
.name("Array Flag")
.required(true)
.description("mark if the input json is array or not")
.addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
.allowableValues("true", "false")
.defaultValue("false")
.build();
/**
* 成功標識.
*/
public static final Relationship SUCCESS = new Relationship.Builder()
.name("SUCCESS")
.description("Succes relationship")
.build();
@Override
public Set<Relationship> getRelationships(){
return relationships;
}
@Override
public List<PropertyDescriptor> getSupportedPropertyDescriptors(){
return properties;
}
/**
* 初始化配置
* @param context
*/
@Override
public void init(final ProcessorInitializationContext context){
List<PropertyDescriptor> properties = new ArrayList<>();
properties.add(JSON_PATH);
properties.add(ARRAY_FLAG);
this.properties = Collections.unmodifiableList(properties);
Set<Relationship> relationships = new HashSet<>();
relationships.add(SUCCESS);
this.relationships = Collections.unmodifiableSet(relationships);
}
@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
final AtomicReference<String> value = new AtomicReference<>();
FlowFile flowfile = session.get();
session.read(flowfile, new InputStreamCallback() {
@Override
public void process(InputStream in) throws IOException {
try{
String json =IOUtils.toString(in, StandardCharsets.UTF_8.name());
String flag = context.getProperty(ARRAY_FLAG).getValue();
if (flag.equalsIgnoreCase(arrayFlag)){
List<Object> dataList = JsonPath.read(json, context.getProperty(JSON_PATH).getValue());
if (ObjectUtils.allNotNull(dataList)){
StringBuilder all = new StringBuilder("[");
int total = 0;
for (Object object : dataList) {
LinkedHashMap<String,Object> dataMap = (LinkedHashMap<String, Object>) object;
Set<String> keys = dataMap.keySet();
int count = 0;
StringBuilder builder = new StringBuilder("{");
for (String key :keys ) {
if (count==keys.size()-1){
builder.append("\""+key+"\":\""+dataMap.get(key)+"\"");
}else{
builder.append("\""+key+"\":\""+dataMap.get(key)+"\",");
}
count++;
}
if (total==dataList.size()-1){
builder.append("}");
}else {
builder.append("},");
}
total++;
all.append(builder.toString());
builder.reverse();
}
all.append("]");
value.set(all.toString());
}
}else {
String result = JsonPath.read(json, context.getProperty(JSON_PATH).getValue());
value.set(result);
}
}catch(Exception ex){
ex.printStackTrace();
getLogger().error("failed to read json string.");
}
}
});
//Write the results to an attribute
String results = value.get();
if(results != null && !results.isEmpty()){
String flag = context.getProperty(ARRAY_FLAG).getValue();
if (flag.equalsIgnoreCase(arrayFlag)){
Map<String,String> data=new HashMap<>(16);
data.put(NiFiConstant.MATCH_ATTR,value.toString());
flowfile = session.putAllAttributes(flowfile,data);
}else {
flowfile = session.putAttribute(flowfile, NiFiConstant.MATCH_ATTR, results);
}
}
//To write the results back out ot flow file
flowfile = session.write(flowfile, new OutputStreamCallback() {
@Override
public void process(OutputStream out) throws IOException {
out.write(value.get().getBytes());
}
});
session.transfer(flowfile, SUCCESS);
}
}
要指明是不是一個json array ，因爲jsonobject 和jsonarray 的解析和接收對象是不一樣的。
3.使用PutFile
指定處理完成之後,文件的輸出地址:
4.整個的流程圖如下：
直接上代碼，按照圖所示來操作，就可以看見對應的文件輸出到目錄裏了。
二、合併文本的內容
假定我們需要在把一個文本內容拼接上另外一個文本內容，那麼在nifi的架子上，該如何操作吶？共分爲三步：
1.Json文檔的輸入端
需要一個GenerateFlowFile處理器,做如下配置:
Custom text 存放的是操作的文本內容，如下所示: 你是哪個 ？
2.拖拽自己實現的處理器
根據要求配置好自己的input value的值，這樣就可以將a中的文本內容進行拼接:
代碼實現如下：
@Tags({"second-example:Combine two sentences!"})
@SeeAlso({})
@SideEffectFree
@CapabilityDescription("merge two content to one together")
@ReadsAttributes({@ReadsAttribute(attribute="", description="")})
@WritesAttributes({@WritesAttribute(attribute="", description="")})
public class SecondProcessor extends AbstractProcessor {
/**
* 屬性描述對象集合
*/
private List<PropertyDescriptor> descriptors;
/**
* 關聯關係集合
*/
private Set<Relationship> relationships;
/**
* 文件設置.
*/
private static final String FILE_NAME = "out-";
private static final String FILE_SUFFIX = ".txt";
public static final PropertyDescriptor INPUT_VALUE = new PropertyDescriptor.Builder()
.name("INPUT_VALUE")
.displayName("INPUT VALUE")
.description("input value for operating")
.required(true)
//非空驗證
.addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
.build();
public static final Relationship RELATIONSHIP_SUCCESS = new Relationship.Builder()
.name("sucess")
.description("example relationship uccess")
.build();
public static final Relationship RELATIONSHIP_FAILURE = new Relationship.Builder()
.name("failure")
.description("example relationship failure")
.build();
public static final PropertyDescriptor CHARSET = new PropertyDescriptor.Builder()
.name("character-set")
.displayName("Character Set")
.required(true)
.defaultValue("UTF-8")
.addValidator(StandardValidators.CHARACTER_SET_VALIDATOR)
.build();
@Override
protected void init(final ProcessorInitializationContext context) {
final List<PropertyDescriptor> descriptors = new ArrayList<PropertyDescriptor>();
descriptors.add(INPUT_VALUE);
descriptors.add(CHARSET);
this.descriptors = Collections.unmodifiableList(descriptors);
final Set<Relationship> relationships = new HashSet<Relationship>();
relationships.add(RELATIONSHIP_SUCCESS);
relationships.add(RELATIONSHIP_FAILURE);
this.relationships = Collections.unmodifiableSet(relationships);
}
@Override
public Set<Relationship> getRelationships() {
return this.relationships;
}
@Override
public final List<PropertyDescriptor> getSupportedPropertyDescriptors() {
return descriptors;
}
@OnScheduled
public void onScheduled(final ProcessContext context) {
getLogger().info("Processor-Name"+context.getName());
Map<PropertyDescriptor, String> dataMap = context.getProperties();
for (Map.Entry<PropertyDescriptor, String> entry : dataMap.entrySet()) {
getLogger().info("key="+entry.getKey().toString()+",value="+entry.getValue());
}
}
@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
FlowFile flowFile = session.get();
if ( flowFile == null ) {
return;
}
final AtomicReference<String> value = new AtomicReference<>();
session.read(flowFile, new InputStreamCallback() {
@Override
public void process(InputStream inputStream) throws IOException {
try{
String inputVal = IOUtils.toString(inputStream, StandardCharsets.UTF_8.name());
//utf8 的設置
final Charset charset = Charset.forName(context.getProperty(CHARSET).getValue());
getLogger().info("得到字符集結果是:"+charset.name());
String current = new String(context.getProperty(INPUT_VALUE).getValue().getBytes(charset),StandardCharsets.UTF_8.name());
String result = "處理結果:" + inputVal + current;
//以 utf8 的方式把流信息寫出去.
getLogger().info("處理得到的結果是:"+result);
value.set(result);
}catch(Exception ex){
ex.printStackTrace();
getLogger().error("failed to read input string!");
}
}
});
String results = value.get();
if(results != null && !results.isEmpty()){
flowFile = session.putAttribute(flowFile, NiFiConstant.MATCH_ATTR, results);
}
//寫入文件信息.
flowFile = session.write(flowFile, new OutputStreamCallback() {
@Override
public void process(OutputStream outputStream) throws IOException {
getLogger().info("寫出的消息是:"+value.get());
byte[] content = value.get().getBytes();
//遠程的輸出流
outputStream.write(content);
//重新定義本地輸出流.
outputStream = new FileOutputStream(new File(FILE_NAME+uuid()+FILE_SUFFIX));
outputStream.write(content);
}
});
session.transfer(flowFile, RELATIONSHIP_SUCCESS);
}
/**
* 產生一個32位的GUID
* @return
*/
public String uuid() {
return getIdentifier().replace("-", "").toUpperCase();
}
}
3.使用PutFile
指定處理完成之後,文件的輸出地址:.
4.整體的流程圖如下所示:
如此就可以完成兩個文本內容的拼接並輸出的操作。
三、給一個文件內容添加頭信息
假定我們需要在把一個文件內部的內容拼接上另外一個文本內容，那麼在nifi的架子上，該如何操作吶？共分爲三步：
1.Json文檔的輸入端
需要一個GenerateFlowFile處理器,做如下配置:
注意file size 和字符集的設置。Custom text 存放的是操作的文本內容，如下所示:你是誰？
2.拖拽自己實現的處理器
根據要求配置好自己的文件的絕對路徑的值，這樣就可以將a中的文本內容進行拼接
需要一個絕對的文件路徑，具體代碼如下：
@Tags({"third-example:deal with content!"})
@SeeAlso({})
@SideEffectFree
@CapabilityDescription("add prefix to given content.")
@ReadsAttributes({@ReadsAttribute(attribute="", description="")})
@WritesAttributes({@WritesAttribute(attribute="", description="")})
public class ThirdProcessor extends AbstractProcessor {
/**
* 屬性描述對象集合
*/
private List<PropertyDescriptor> descriptors;
/**
* 關聯關係集合
*/
private Set<Relationship> relationships;
/**
* 文件設置.
*/
private static final String FILE_NAME = "combine-";
private static final String FILE_SUFFIX = ".txt";
public static final PropertyDescriptor ABSOLUTE_PATH = new PropertyDescriptor.Builder()
.name("ABSOLUTE_PATH")
.displayName("ABSOLUT PATH")
.description("input file path for operating")
.required(true)
//非空驗證
.addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
.build();
public static final Relationship SHIP_SUCCESS = new Relationship.Builder()
.name("sucess")
.description("example relationship uccess")
.build();
public static final Relationship SHIP_FAILURE = new Relationship.Builder()
.name("failure")
.description("example relationship failure")
.build();
public static final PropertyDescriptor CHARSET = new PropertyDescriptor.Builder()
.name("character-set")
.displayName("Character Set")
.required(true)
.defaultValue("UTF-8")
.addValidator(StandardValidators.CHARACTER_SET_VALIDATOR)
.build();
@Override
protected void init(final ProcessorInitializationContext context) {
final List<PropertyDescriptor> descriptors = new ArrayList<PropertyDescriptor>();
descriptors.add(ABSOLUTE_PATH);
descriptors.add(CHARSET);
this.descriptors = Collections.unmodifiableList(descriptors);
final Set<Relationship> ships = new HashSet<Relationship>();
ships.add(SHIP_SUCCESS);
ships.add(SHIP_FAILURE);
this.relationships = Collections.unmodifiableSet(ships);
}
@Override
public Set<Relationship> getRelationships() {
return this.relationships;
}
@Override
public final List<PropertyDescriptor> getSupportedPropertyDescriptors() {
return descriptors;
}
@OnScheduled
public void onScheduled(final ProcessContext context) {
getLogger().info("Processor-Name"+context.getName());
Map<PropertyDescriptor, String> dataMap = context.getProperties();
for (Map.Entry<PropertyDescriptor, String> entry : dataMap.entrySet()) {
getLogger().info("key="+entry.getKey().toString()+",value="+entry.getValue());
}
}
@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
FlowFile flowFile = session.get();
if ( flowFile == null ) {
return;
}
final AtomicReference<String> value = new AtomicReference<>();
//utf8 的設置
final Charset charset = Charset.forName(context.getProperty(CHARSET).getValue());
session.read(flowFile, new InputStreamCallback() {
@Override
public void process(InputStream inputStream) throws IOException {
try{
String headerDesc = IOUtils.toString(inputStream, StandardCharsets.UTF_8.name());
String filePath = context.getProperty(ABSOLUTE_PATH).getValue();
InputStreamReader inRd = new InputStreamReader(new FileInputStream(filePath),charset);
BufferedReader reader = new BufferedReader(inRd);
String line=null;
StringBuilder builder = new StringBuilder();
while (null!=(line=reader.readLine())){
getLogger().info("文件信息是:"+line);
builder.append(headerDesc+new String(line.getBytes(charset),StandardCharsets.UTF_8.name())+"\n\t");
}
//以 utf8 的方式把流信息寫出去.
getLogger().info("處理得到的結果是:"+builder.toString());
value.set(builder.toString());
}catch(Exception ex){
ex.printStackTrace();
getLogger().error("failed to read input string!");
}
}
});
String results = value.get();
if(results != null && !results.isEmpty()){
flowFile = session.putAttribute(flowFile, NiFiConstant.MATCH_ATTR, results);
}
//寫入文件信息.
flowFile = session.write(flowFile, new OutputStreamCallback() {
@Override
public void process(OutputStream outputStream) throws IOException {
getLogger().info("寫出的消息是:"+value.get());
byte[] content = value.get().getBytes();
//遠程的輸出流
outputStream.write(content);
//重新定義本地輸出流.
outputStream = new FileOutputStream(new File(FILE_NAME+uuid()+FILE_SUFFIX));
outputStream.write(content);
}
});
session.transfer(flowFile, SHIP_SUCCESS);
}
/**
* 產生一個32位的GUID
* @return
*/
public String uuid() {
return getIdentifier().replace("-", "").toUpperCase();
}
}
3.使用PutFile
指定處理完成之後,文件的輸出地址
4.整體的流程圖如下所示
如此挨個執行，不報錯的情況下，就可以看見執行的結果了。
本文只是一個簡單描述下如何基於nifi框架，來實現自己的業務邏輯，下一篇我就複雜使用下，看看如何操作。
如有不明白的，請微信搜索公衆號 ：codingba ,我會一一解答。