File: ReusingReducerLogic.md

package info (click to toggle)
node-redux 4.2.1-1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm, forky, sid, trixie
  • size: 9,156 kB
  • sloc: javascript: 7,634; makefile: 2
file content (140 lines) | stat: -rw-r--r-- 5,879 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
---
id: reusing-reducer-logic
title: Reusing Reducer Logic
sidebar_label: Reusing Reducer Logic
hide_title: true
---

# Reusing Reducer Logic

As an application grows, common patterns in reducer logic will start to emerge. You may find several parts of your reducer logic doing the same kinds of work for different types of data, and want to reduce duplication by reusing the same common logic for each data type. Or, you may want to have multiple "instances" of a certain type of data being handled in the store. However, the global structure of a Redux store comes with some trade-offs: it makes it easy to track the overall state of an application, but can also make it harder to "target" actions that need to update a specific piece of state, particularly if you are using `combineReducers`.

As an example, let's say that we want to track multiple counters in our application, named A, B, and C. We define our initial `counter` reducer, and we use `combineReducers` to set up our state:

```js
function counter(state = 0, action) {
  switch (action.type) {
    case 'INCREMENT':
      return state + 1
    case 'DECREMENT':
      return state - 1
    default:
      return state
  }
}

const rootReducer = combineReducers({
  counterA: counter,
  counterB: counter,
  counterC: counter,
})
```

Unfortunately, this setup has a problem. Because `combineReducers` will call each slice reducer with the same action, dispatching `{type : 'INCREMENT'}` will actually cause _all three_ counter values to be incremented, not just one of them. We need some way to wrap the `counter` logic so that we can ensure that only the counter we care about is updated.

## Customizing Behavior with Higher-Order Reducers

As defined in [Splitting Reducer Logic](SplittingReducerLogic.md), a _higher-order reducer_ is a function that takes a reducer function as an argument, and/or returns a new reducer function as a result. It can also be viewed as a "reducer factory". `combineReducers` is one example of a higher-order reducer. We can use this pattern to create specialized versions of our own reducer functions, with each version only responding to specific actions.

The two most common ways to specialize a reducer are to generate new action constants with a given prefix or suffix, or to attach additional info inside the action object. Here's what those might look like:

```js
function createCounterWithNamedType(counterName = '') {
  return function counter(state = 0, action) {
    switch (action.type) {
      case `INCREMENT_${counterName}`:
        return state + 1
      case `DECREMENT_${counterName}`:
        return state - 1
      default:
        return state
    }
  }
}

function createCounterWithNameData(counterName = '') {
  return function counter(state = 0, action) {
    const { name } = action
    if (name !== counterName) return state

    switch (action.type) {
      case `INCREMENT`:
        return state + 1
      case `DECREMENT`:
        return state - 1
      default:
        return state
    }
  }
}
```

We should now be able to use either of these to generate our specialized counter reducers, and then dispatch actions that will affect the portion of the state that we care about:

```js
const rootReducer = combineReducers({
  counterA: createCounterWithNamedType('A'),
  counterB: createCounterWithNamedType('B'),
  counterC: createCounterWithNamedType('C'),
})

store.dispatch({ type: 'INCREMENT_B' })
console.log(store.getState())
// {counterA : 0, counterB : 1, counterC : 0}
```

We could also vary the approach somewhat, and create a more generic higher-order reducer that accepts both a given reducer function and a name or identifier:

```js
function counter(state = 0, action) {
  switch (action.type) {
    case 'INCREMENT':
      return state + 1
    case 'DECREMENT':
      return state - 1
    default:
      return state
  }
}

function createNamedWrapperReducer(reducerFunction, reducerName) {
  return (state, action) => {
    const { name } = action
    const isInitializationCall = state === undefined
    if (name !== reducerName && !isInitializationCall) return state

    return reducerFunction(state, action)
  }
}

const rootReducer = combineReducers({
  counterA: createNamedWrapperReducer(counter, 'A'),
  counterB: createNamedWrapperReducer(counter, 'B'),
  counterC: createNamedWrapperReducer(counter, 'C'),
})
```

You could even go as far as to make a generic filtering higher-order reducer:

```js
function createFilteredReducer(reducerFunction, reducerPredicate) {
    return (state, action) => {
        const isInitializationCall = state === undefined;
        const shouldRunWrappedReducer = reducerPredicate(action) || isInitializationCall;
        return shouldRunWrappedReducer ? reducerFunction(state, action) : state;
    }
}

const rootReducer = combineReducers({
    // check for suffixed strings
    counterA : createFilteredReducer(counter, action => action.type.endsWith('_A')),
    // check for extra data in the action
    counterB : createFilteredReducer(counter, action => action.name === 'B'),
    // respond to all 'INCREMENT' actions, but never 'DECREMENT'
    counterC : createFilteredReducer(counter, action => action.type === 'INCREMENT')
};
```

These basic patterns allow you to do things like having multiple instances of a smart connected component within the UI, or reuse common logic for generic capabilities such as pagination or sorting.

In addition to generating reducers this way, you might also want to generate action creators using the same approach, and could generate them both at the same time with helper functions. See [Action/Reducer Generators](https://github.com/markerikson/redux-ecosystem-links/blob/master/action-reducer-generators.md) and [Reducers](https://github.com/markerikson/redux-ecosystem-links/blob/master/reducers.md) libraries for action/reducer utilities.