2
votes

I am building a REST API in NodeJS. I am building the server-side pagination, sorting and filtering.

I have a simple implementation of the filter method. This works if the item does not contain any empty strings. However if the item contains an empty string, the .toString() method fails because item[filterBy] is null.

Cannot read property 'toString' of null

How can I update the filter function to simply return true if the item contains an empty string, to skip the comparison operations?

  // Apply filter to array of objects
  // TODO: Empty strings breaks the filter
  items = items.filter(item =>
    item[filterBy].toString().trim().toLowerCase().indexOf(filterValue.toLowerCase()) !== -1)
5

5 Answers

2
votes

To skip the item contains an empty string, try this:

item[filterBy] && item[filterBy].toString()

So, your code will be:

items = items.filter(item =>
item[filterBy] && item[filterBy].toString().trim().toLowerCase().indexOf(filterValue.toLowerCase()) !== -1)
2
votes

Perhaps something like this:

const filterValue = 'abc'
const filterBy = 'name'
const items = [
  {x: 1, name: 'abc'},
  {x: 2, name: 'def'},
  {x: 3, noname: 'def'},
  {x: 4, name: 'ABCDEF'},
]

const newItems = items.filter(item =>
    !(filterBy in item) || item[filterBy].toString().trim().toLowerCase().indexOf(filterValue.toLowerCase()) !== -1)
    
console.log(newItems)

All we do here is check if the item has a filterBy property.

It's not quite clear to me what the issue is with empty strings. It seems most likely that you would get that error if item[filterBy] is undefined. That's what we check here.

If instead you want to skip those that don't have the relevant property, switch from !(filterBy in item) || ... to (filterBy in item) && ....

1
votes

Maybe:

items = items.filter(
    item => item[filterBy] && item[filterBy].toString ?
        item[filterBy].toString().trim().toLowerCase().indexOf(filterValue.toLowerCase()) !== -1 :
        false
)
1
votes

flatMap() is .map() and .flat() combined. It can act like a reverse .filter() by directly returning or not returning values. See demo for details.

const items = [{
  id: 'a',
  pn: 'B',
  desc: ''
}, {
  id: 'b',
  pn: 'c',
  desc: 'd'
}, {
  id: 'c',
  pn: 'k',
  desc: null
}, {
  id: 'i',
  pn: 'k',
  desc: 2
},, {
  id: 'o',
  pn: 'x',
  desc: 3
}];

// Utility function that returns a 'friendlier' value
function clean(obj) {
  return obj == null ? false : Array.isArray(obj) ? obj : obj.constructor.name.toLowerCase() === "object" ? obj.toString().trim().toLowerCase() : typeof obj === "string" ? obj.toString().trim().toLowerCase() : Number(parseFloat(obj)) === obj ? obj : false;
}

// Filters by a single key and multiple values
let filterKey = clean('desc');
let filterVal = clean(['d', 2, 'f']);

/* It returns each value in a sub-array an empty array will remove
the value. In the final stage it flattens the array of arrays into a normal array
*/
let list = items.flatMap(item => {
  return Object.entries(item).flatMap(([key, value]) => filterKey === (clean(key)) && [...filterVal].includes(clean(value)) ? [clean(value)] :[]);
});

console.log(list);
0
votes

It is correct to chain filters and it does not affect performance.

const filterValue = 'John'
const filterBy = 'name'
const items = [
  {id: 1, name: 'John'},
  {id: 2, name: 'Doe, John'},
  {id: 3, ref: 1},
  {id: 4, name: 'No one'},
  {id: 5, name: null}
]

let fItems = items.filter(item => item[filterBy])
                  .filter(item => item[filterBy].toString().trim().toLowerCase()
                     .indexOf(filterValue.toLowerCase()) !== -1)
   
console.log(fItems);

Update: fixed the code.