对于这个需求，我实现并发布了一个通用组件。看到

https://www.npmjs.com/package/w-ng5

在使用此组件之前，使用npm安装此包:

npm install w-ng5 --save

之后，在app.module中导入模块

...
import { PipesModule } from 'w-ng5';

下一步，在app.module中添加declare部分:

imports: [
  PipesModule,
  ...
]

样本用

过滤简单字符串

<input type="text"  [(ngModel)]="filtroString">
<ul>
  <li *ngFor="let s of getStrings() | filter:filtroString">
    {{s}}
  </li>
</ul>

在级别2中过滤复杂字符串字段“Value”

<input type="text"  [(ngModel)]="search">
<ul>
  <li *ngFor="let s of getComplexTypesExtends() | filter:[{field:'n1.n2.valor2', value: search}]">
    {{s.nome}} - {{s.idade}} - {{s.n1.valor1}} - {{s.n1.n2.valor2}}
  </li>
</ul>

过滤复杂字符串-中间字段- 1级中的“值”

<input type="text"  [(ngModel)]="search3">
<ul>
  <li *ngFor="let s of getComplexTypesExtends() | filter:[{field:'n1.valor1', value: search3}]">
    {{s.nome}} - {{s.idade}} - {{s.n1.valor1}} - {{s.n1.n2.valor2}}
  </li>
</ul>

过滤复杂数组的简单字段“Nome”级别为0

<input type="text"  [(ngModel)]="search2">
<ul>
  <li *ngFor="let s of getComplexTypesExtends() | filter:[{field:'nome', value: search2}]">
    {{s.nome}} - {{s.idade}} - {{s.n1.valor1}} - {{s.n1.n2.valor2}}
  </li>
</ul>

在树字段中过滤-在第2级中字段'Valor'或在第1级中字段'Valor'或在第0级中字段'Nome'

<input type="text"  [(ngModel)]="search5">
<ul>
  <li *ngFor="let s of getComplexTypesExtends() | filter:[{field:'n1.n2.valor2', value: search5}, {field:'n1.valor1', value: search5}, {field:'nome', value: search5}]">
    {{s.nome}} - {{s.idade}} - {{s.n1.valor1}} - {{s.n1.n2.valor2}}
  </li>
</ul>

过滤不存在的字段- 'Valor'在不存在的级别3

<input type="text"  [(ngModel)]="search4">
<ul>
  <li *ngFor="let s of getComplexTypesExtends() | filter:[{field:'n1.n2.n3.valor3', value: search4}]">
    {{s.nome}} - {{s.idade}} - {{s.n1.valor1}} - {{s.n1.n2.valor2}}
  </li>
</ul>

该组件具有无限属性级别…

第一步是在组件中使用@Pipe创建Filter。ts文件:

your.component.ts

import { Component, Pipe, PipeTransform, Injectable } from '@angular/core';
import { Person} from "yourPath";

@Pipe({
  name: 'searchfilter'
})
@Injectable()
export class SearchFilterPipe implements PipeTransform {
  transform(items: Person[], value: string): any[] {
    if (!items || !value) {
      return items;
    }
    console.log("your search token = "+value);
    return items.filter(e => e.firstName.toLowerCase().includes(value.toLocaleLowerCase()));
  }
}
@Component({
  ....
    persons;

    ngOnInit() {
         //inicial persons arrays
    }
})

和Person对象的数据结构:

person.ts

export class Person{
    constructor(
        public firstName: string,
        public lastName: string
    ) { }
}

在你的html文件视图中:

your.component.html

    <input class="form-control" placeholder="Search" id="search" type="text" [(ngModel)]="searchText"/>
    <table class="table table-striped table-hover">
      <colgroup>
        <col span="1" style="width: 50%;">
        <col span="1" style="width: 50%;">
      </colgroup>
      <thead>
        <tr>
          <th>First name</th>
          <th>Last name</th>
        </tr>
      </thead>
      <tbody>
        <tr *ngFor="let person of persons | searchfilter:searchText">
          <td>{{person.firstName}}</td>
          <td>{{person.lastName}}</td>
        </tr>
      </tbody>
    </table>

这是你的数组

products: any = [
        {
            "name": "John-Cena",
                    },
        {
            "name": "Brock-Lensar",

        }
    ];

这是你的ngFor循环 过滤者:

<input type="text" [(ngModel)]='filterText' />
    <ul *ngFor='let product of filterProduct'>
      <li>{{product.name }}</li>
    </ul>

这里我使用的是产品的filterProduct instant，因为我想保留原始数据。 这里model _filterText用作输入框。当有任何变化setter函数将调用。 在setFilterText调用performProduct时，它只返回与输入匹配的结果。我用小写表示不区分大小写。

filterProduct = this.products;
_filterText : string;
    get filterText() : string {
        return this._filterText;
    }

    set filterText(value : string) {
        this._filterText = value;
        this.filterProduct = this._filterText ? this.performProduct(this._filterText) : this.products;

    } 

    performProduct(value : string ) : any {
            value = value.toLocaleLowerCase();
            return this.products.filter(( products : any ) => 
                products.name.toLocaleLowerCase().indexOf(value) !== -1);
        }

我正在寻找一些东西，让一个过滤器传递一个对象，然后我可以使用它像多过滤器:

我做了这个美容方案:

filter.pipe.ts

import { PipeTransform, Pipe } from '@angular/core';

@Pipe({
  name: 'filterx',
  pure: false
})
export class FilterPipe implements PipeTransform {
 transform(items: any, filter: any, isAnd: boolean): any {
  let filterx=JSON.parse(JSON.stringify(filter));
  for (var prop in filterx) {
    if (Object.prototype.hasOwnProperty.call(filterx, prop)) {
       if(filterx[prop]=='')
       {
         delete filterx[prop];
       }
    }
 }
if (!items || !filterx) {
  return items;
}

return items.filter(function(obj) {
  return Object.keys(filterx).every(function(c) {
    return obj[c].toLowerCase().indexOf(filterx[c].toLowerCase()) !== -1
  });
  });
  }
}

component.ts

slotFilter:any={start:'',practitionerCodeDisplay:'',practitionerName:''};

componet.html

             <tr>
                <th class="text-center">  <input type="text" [(ngModel)]="slotFilter.start"></th>
                <th class="text-center"><input type="text" [(ngModel)]="slotFilter.practitionerCodeDisplay"></th>
                <th class="text-left"><input type="text" [(ngModel)]="slotFilter.practitionerName"></th>
                <th></th>
              </tr>


 <tbody *ngFor="let item of practionerRoleList | filterx: slotFilter">...

管子是最好的方法。但低于1也可以。

<div *ng-for="#item of itemsList">
  <ng-container *ng-if="conditon(item)">
    // my code
  </ng-container>
</div>

Angular2中的管道类似于命令行中的管道。前面每个值的输出在管道之后被送入过滤器，这使得链过滤器很容易，就像这样:

<template *ngFor="#item of itemsList">
    <div *ngIf="conditon(item)">{item | filter1 | filter2}</div>
</template>