Related
I've got a simple problem, but I'm struggling to find the easiest solution without transforming the array a hundred times.
I want to do a simple stacked graph in google sheets, with weeks on X and values on Y. I got the values for each week, but only for weeks, that have a value.
The values are all calculations I've done with google apps script/ js.
person1 = [[2019/37,2], [2019/42,3]] and so on, for multiple persons and for 80 weeks in total.
The num value is the total value after each week. So I want the array to be filled up with the missing weeks. Therefore I mapped this to another array, where I have all the weeks but no values, giving these weeks the value 0:
person1= [[2019/37,2],[2019/38,0],[2019/39,0],...,[2019/42,3],[2019/43,0],[2019/44,0],...]
This of course does not fit to see a progress in the graph.
So I need something to set the weeks, which were filled up, to the previous value, resulting in
person1= [[2019/37,2],[2019/38,2],[2019/39,2],...,[2019/42,3],[2019/43,3],[2019/44,3],...]
Looping through this and setting the values with something like person[i][1] == person[i-1][1] seems not to be a good practice of course.
So, what would be the best way to achieve this? I'm kind of stuck with this now, I feel like I don't see the forest for the trees.
Thanks in advance!
code:
let valueArray = [[2019/37,2], [2019/42,3]]
let weeksArray = [2019/38,2019/39,2019/40,2019/41...]
//find missing weeks
let notFound = weeksArray.filter(el => valueArray.includes(el) == false).map(x => [x,0]);
//concat and sort
let outArray = arr.concat(notFound).sort((a,b)=> a[0].localeCompare(b[0]));
//output:
//[[2019/37,2],[2019/38,0],[2019/39,0],...,[2019/42,3],[2019/43,0],[2019/44,0],...]
Solution:
Since you already have the expanded array, you can use map on the whole array and use a function to replace the values:
var weeks = [[2019/37,2],[2019/38,0],[2019/39,0],[2019/40,3],[2019/41,0],[2019/42,4],[2019/43,0],[2019/44,0]];
weeks.map((a,b)=>{weeks[b][1] = (a[1] == 0 && b > 0) ? weeks[b-1][1] : weeks[b][1]});
To make it more readable, this is the same as:
weeks.forEach(function missing(item,index,arr) {
if (item[1] == 0 && index > 0) {
arr[index][1] = arr[index-1][1];
}
}
);
Console log:
References:
Arrow Functions
Conditional Operator
Array.prototype.map()
function fixArray() {
var array = [["2019/1", "1"], ["2019/10", "2"], ["2019/20", "3"], ["2019/30", "4"], ["2019/40", "5"]];
var oA = [];
array.forEach(function (r, i) {
oA.push(r);
let t1 = r[0].split('/');
let diff;
if (i + 1 < array.length) {
let inc = 1;
let t2 = array[i + 1][0].split('/');
if (t1[0] == t2[0] && t2[1] - t1[1] > 1) {
do {
let t3 = ['', ''];
t3[0] = t1[0] + '/' + Number(parseInt(t1[1]) + inc);
t3[1] = r[1];
diff = t2[1] - t1[1] - inc;
oA.push(t3);
inc++;
} while (diff > 1);
}
}
});
let end = "is near";
console.log(JSON.stringify(oA));
}
console.log:
[["2019/1","1"],["2019/2","1"],["2019/3","1"],["2019/4","1"],["2019/5","1"],["2019/6","1"],["2019/7","1"],["2019/8","1"],["2019/9","1"],["2019/10","2"],["2019/11","2"],["2019/12","2"],["2019/13","2"],["2019/14","2"],["2019/15","2"],["2019/16","2"],["2019/17","2"],["2019/18","2"],["2019/19","2"],["2019/20","3"],["2019/21","3"],["2019/22","3"],["2019/23","3"],["2019/24","3"],["2019/25","3"],["2019/26","3"],["2019/27","3"],["2019/28","3"],["2019/29","3"],["2019/30","4"],["2019/31","4"],["2019/32","4"],["2019/33","4"],["2019/34","4"],["2019/35","4"],["2019/36","4"],["2019/37","4"],["2019/38","4"],["2019/39","4"],["2019/40","5"]]
For example, take the time range from 05/10/2019 to 05/25/2019.
Dates on this interval need to be aggregated like this (2019 is omitted for brevity):
const result = [
[ '05-10', '05-11', '05-12'], // week 1
['05-13', '05-14', '05-15', '05-16', '05-17', '05-18', '05-19'], // week 2
['05-20', '05-21', '05-22', '05-23', '05-24', '05-25' ], // week 3
];
What is the best way to solve this problem with JS?
Is it possible to implement this by setting the beginning of the week on any day?
Will packages moment and moment-range help in this?
You can go through the dates, and if the day is 1 (Monday), create a new Array in your results:
const startDate = new Date('05-10-2019'),
endDate = new Date('05-25-2019'),
result = [];
function _twoDigits(x) {
return String(x).padStart(2, '0');
}
let tmp = startDate;
do {
if (tmp.getDay() === 1 || result.length === 0) {
// Create a week Array
result.push([]);
}
const str = `${_twoDigits(tmp.getMonth() + 1)}-${_twoDigits(tmp.getDate())}`;
// Add this date to the last week Array
result[result.length - 1].push(str);
// Add 24 hours
tmp = new Date(tmp.getTime() + 86400000);
} while (tmp.getTime() <= endDate.getTime());
console.log(result);
Note: MomentJS may help, but it's a big library. If you only need to do 2 or 3 basic things with dates, I would recommend not using it. If you need to do a lot of work with dates, then yes, it's a powerful library that will save you a lot of time.
Here is one possible implementation if you are interested in moment.js code.
But as blex said it's a large lib.
const start = moment('2019-05-10');
const end = moment('2019-05-25');
const array = [[]];
const from_date = moment(start).startOf('isoWeek');
const to_date = moment(end).endOf('isoWeek');
let j = 0;
let added = 0;
for (let currentDate = moment(from_date); currentDate <= to_date; currentDate.add(1, 'day')) {
if (added === 7) {
array.push([]);
j++;
added = 0;
}
if (currentDate.isBetween(start, end, null, '[]')) {
array[j].push(currentDate.format('MM-DD'));
}
else {
array[j].push('');
}
added++;
}
document.getElementById('output').innerText = JSON.stringify(array);
<script src="https://cdnjs.cloudflare.com/ajax/libs/moment.js/2.24.0/moment.min.js"></script>
<p id="output"></p>
Ive been reading everything online but its not exactly what I need
var x = 'a1b2c3d4e5'
I need something to get me to
using 1 the answer should be abcde
using 2 the answer should be 12345
using 3 the answer should be b3e
the idea behind it if using 1 it grabs 1 skips 1
the idea behind it if using 2 it grabs 2 skips 2
the idea behind it if using 3 it grabs 3 skips 3
I dont want to use a for loop as it is way to long especially when your x is longer than 300000 chars.
is there a regex I can use or a function that Im not aware of?
update
I'm trying to some how implement your answers but when I use 1 that's when I face the problem. I did mention trying to stay away from for-loops the reason is resources on the server. The more clients connect the slower everything becomes. So far array.filter seem a lot quicker.
As soon as I've found it I'll accept the answer.
As others point out, it's not like regular expressions are magic; there would still be an underlying looping mechanism. Don't worry though, when it comes to loops, 300,000 is nothing -
console.time('while')
let x = 0
while (x++ < 300000)
x += 1
console.timeEnd('while')
// while: 5.135 ms
console.log(x)
// 300000
Make a big string, who cares? 300,000 is nothing -
// 10 chars repeated 30,000 times
const s =
'abcdefghij'.repeat(30000)
console.time('while string')
let x = 0
let interval = 2
let values = []
while (x < s.length)
{ values.push(s[x])
x += interval
}
let result = values.join('')
console.timeEnd('while string')
// while string: 31.990ms
console.log(result)
console.log(result.length)
// acegiacegiacegiacegiacegiacegiacegiacegia...
// 150000
Or use an interval of 3 -
const s =
'abcdefghij'.repeat(30000)
console.time('while string')
let x = 0
let interval = 3
let values = []
while (x < s.length)
{ values.push(s[x])
x += interval
}
let result = values.join('')
console.timeEnd('while string')
// while string: 25.055ms
console.log(result)
console.log(result.length)
// adgjcfibehadgjcfibehadgjcfibehadgjcfibe...
// 100000
Using a larger interval obviously results in fewer loops, so the total execution time is lower. The resulting string is shorter too.
const s =
'abcdefghij'.repeat(30000)
console.time('while string')
let x = 0
let interval = 25 // big interval
let values = []
while (x < s.length)
{ values.push(s[x])
x += interval
}
let result = values.join('')
console.timeEnd('while string')
// while string: 6.130
console.log(result)
console.log(result.length)
// afafafafafafafafafafafafafafafafafafafafafafa...
// 12000
You can achieve functional style and stack-safe speed simultaneously -
const { loop, recur } = require('./lib')
const everyNth = (s, n) =>
loop
( (acc = '', x = 0) =>
x >= s.length
? acc
: recur(acc + s[x], x + n)
)
const s = 'abcdefghij'.repeat(30000)
console.time('loop/recur')
const result = everyNth(s, 2)
console.timeEnd('loop/recur')
// loop/recur: 31.615 ms
console.log(result)
console.log(result.length)
// acegiacegiacegiacegiacegiacegiacegia ...
// 150000
The two are easily implemented -
const recur = (...values) =>
({ recur, values })
const loop = f =>
{ let acc = f()
while (acc && acc.recur === recur)
acc = f(...acc.values)
return acc
}
// ...
module.exports =
{ loop, recur, ... }
And unlike the [...str].filter(...) solutions which will always iterate through every element, our custom loop is much more flexible and receives speed benefit when a higher interval n is used -
console.time('loop/recur')
const result = everyNth(s, 25)
console.timeEnd('loop/recur')
// loop/recur: 5.770ms
console.log(result)
console.log(result.length)
// afafafafafafafafafafafafafafa...
// 12000
const recur = (...values) =>
({ recur, values })
const loop = f =>
{ let acc = f()
while (acc && acc.recur === recur)
acc = f(...acc.values)
return acc
}
const everyNth = (s, n) =>
loop
( (acc = '', x = 0) =>
x >= s.length
? acc
: recur(acc + s[x], x + n)
)
const s = 'abcdefghij'.repeat(30000)
console.time('loop/recur')
const result = everyNth(s, 2)
console.timeEnd('loop/recur')
// loop/recur: 31.615 ms
console.log(result)
console.log(result.length)
// acegiacegiacegiacegiacegiacegiacegia ...
// 150000
Since I'm not an expert of regex, I'd use some fancy es6 functions to filter your chars.
var x = 'a1b2c3d4e5'
var n = 2;
var result = [...x].filter((char, index) => index % n == 0);
console.log(result);
Note that because 0 % 2 will also return 0, this will always return the first char. You can filter the first char by adding another simple check.
var result = [...x].filter((char, index) => index > 0 && index % n == 0);
As a variant:
function getNth(str, nth) {
return [...str].filter((_, i) => (i + 1) % nth === 0).join('');
}
console.log(getNth('a1b2c3d4e5', 2)); // 12345
console.log(getNth('a1b2c3d4e5', 3)); // b3e
What I'd suggest, to avoid having to iterate over the entire array, is to step straight into the known nth's.
Here's a couple of flavors:
function nthCharSubstr(str, nth) {
let res = "";
for (let i = nth - 1; i < str.length; i += nth) {
res += string[i];
}
return res;
}
More ES6-y:
const nthCharSubstr = (str, nth) =>
[...Array(parseInt(str.length / nth)).keys()] // find out the resulting number of characters and create and array with the exact length
.map(i => nth + i * nth - 1) // each item in the array now represents the resulting character's index
.reduce((res, i) => res + str[i], ""); // pull out each exact character and group them in a final string
This solution considers this comment as being valid.
I have this function:
function getInfoSchoolTime() {
var date = new Date();
var schoolBellTime = ["8:10","9:02","9:54","9:59","10:51","11:43","11:58","12:48","13:35","13:40","14:10","15:02","15:54"];
var remainingTime, currentHour;
for (var i = 0; i < schoolBellTime.length-1; i++) {
var startTime = schoolBellTime[i].split(":");
var endTime = schoolBellTime[i+1].split(":");
if (parseInt(startTime[0]) >= date.getHours() && parseInt(startTime[1]) >= date.getMinutes())
if (parseInt(endTime[0]) <= date.getHours() && parseInt(endTime[1]) <= date.getMinutes()) {
currentHour = i;
remainingTime=(parseInt(endTime[1])-date.getMinutes()+60)%60;
break;
}
}
if (currentHour == undefined)
return {current: -1, remaining: "not available"};
return {current: currentHour, remaining: remainingTime};
}
var info = getInfoSchoolTime();
console.log(info.current, info.remaining);
I have the schoolBellTime array that contains the timestamps of my school bell (I know, my school has strange bell times, these timestamps includes playtimes and lunchtime), this function is meant to return the 1st hour/2nd hour/3rd hour ... and the minutes that remains to the next hour/breaktime.
I checked all the code and can't find the error, it keeps returning {current: -1, remaining: "not available"}
The function at the top: setDateTime() takes a date and a time, and constructs a date object for that time.
Then I updated your function, I convert start and end to times on the current day, and then check if date.getTime() occurs between them. Then I simply subtract date.getTime() from end, and convert the result to minutes from milliseconds.
var setDateTime = function(date, str) {
var sp = str.split(':');
date.setHours(parseInt(sp[0], 10));
date.setMinutes(parseInt(sp[1], 10));
return date;
}
function getInfoSchoolTime() {
var date = new Date();
var schoolBellTime = ["8:10", "9:02", "9:54", "9:59", "10:51", "11:43", "11:58", "12:48", "13:35", "13:40", "14:10", "14:10", "15:02", "15:54"];
var remainingTime, currentHour, currentPeriod;
for (var i = 0; i < schoolBellTime.length - 1; i++) {
start = setDateTime(new Date(), schoolBellTime[i])
end = setDateTime(new Date(), schoolBellTime[i + 1])
if (date.getTime() > start.getTime() && date.getTime() < end.getTime()) {
currentHour = i
remainingTime = end.getTime() - date.getTime()
currentPeriod = ([schoolBellTime[i], schoolBellTime[i+1]]).join('-')
}
}
return {current: currentHour, currentPeriod: currentPeriod, remaining: Math.round(remainingTime * 0.0000166667)}
}
console.log(getInfoSchoolTime())
Here's a somewhat different approach, both to the code and the API. It uses two helper functions. Each should be obvious with a single example: pad(7) //=> "07" and pairs(['foo', 'bar', 'baz', 'qux']) //=> [['foo', 'bar'], ['bar', 'baz'], ['baz', 'qux']].
The main function takes a list of bell times and returns a function which itself accepts a date object and returns the sort of output you're looking for (period, remaining time in period.) This API makes it much easier to test.
const pad = nbr => ('00' + nbr).slice(-2)
const pairs = vals => vals.reduce((res, val, idx) => idx < 1 ? res : res.concat([[vals[idx - 1], val]]), [])
const schoolPeriods = (schoolBellTime) => {
const subtractTimes = (t1, t2) => 60 * t1.hour + t1.minute - (60 * t2.hour + t2.minute)
const periods = pairs(schoolBellTime.map(time => ({hour: time.split(':')[0], minute: +time.split(':')[1]})))
return date => {
const current = {hour: date.getHours(), minute: date.getMinutes()}
if (subtractTimes(current, periods[0][0]) < 0) {
return {message: 'before school day'}
}
if (subtractTimes(current, periods[periods.length - 1][1]) > 0) {
return {message: 'after school day'}
}
const idx = periods.findIndex(period => subtractTimes(current, period[0]) >= 0 && subtractTimes(period[1], current) > 0)
const period = periods[idx]
return {
current: idx + 1,
currentPeriod: `${period[0].hour}:${pad(period[0].minute)} - ${period[1].hour}:${pad(period[1].minute)}`,
remaining: subtractTimes(period[1], current)
}
}
}
const getPeriod = schoolPeriods(["8:10","9:02","9:54","9:59","10:51","11:43","11:58","12:48","13:35","13:40","14:10","14:10","15:02","15:54"])
console.log("Using current time")
console.log(getPeriod(new Date()))
console.log("Using a fixed time")
console.log(getPeriod(new Date(2017, 11, 22, 14, 27))) // will Christmas break ever come?!
I made a random guess at the behavior you would want if the date is outside the period range.
Internally, it creates a list of period objects that look like
[{hour:9, minute: 59}, {hour: 10, minute: 51}]
Perhaps it would be cleaner if instead of a two-element array it was an object with start and end properties. That would be an easy change.
Do note that for this to make sense, the bells need to be listed in order. We could fix this with a sort call, but I don't see a good reason to do so.
Here is an ES6 example using deconstruct (const [a,b]=[1,2]), array map, array reduce, partial application (closure) and fat arrow function syntax.
This may not work in older browsers.
//pass date and bellTimes to function so you can test it more easily
// you can partially apply bellTimes
const getInfoSchoolTime = bellTimes => {
//convert hour and minute to a number
const convertedBellTimes = bellTimes
.map(bellTime=>bellTime.split(":"))//split hour and minute
.map(([hour,minute])=>[new Number(hour),new Number(minute)])//convert to number
.map(([hour,minute])=>(hour*60)+minute)//create single number (hour*60)+minutes
.reduce(//zip with next
(ret,item,index,all)=>
(index!==all.length-1)//do not do last one, create [1,2][2,3][3,4]...
? ret.concat([[item,all[index+1]]])
: ret,
[]
);
return date =>{
//convert passed in date to a number (hour*60)+minutes
const passedInTime = (date.getHours()*60)+date.getMinutes();
return convertedBellTimes.reduce(
([ret,goOn],[low,high],index,all)=>
//if goOn is true and passedInTime between current and next bell item
(goOn && passedInTime<high && passedInTime>=low)
? [//found the item, return object and set goOn to false
{
current: index+1,
currentPeriod: bellTimes[index]+"-"+bellTimes[index+1],
remaining: high-passedInTime
},
false//set goOn to false, do not continue checking
]
: [ret,goOn],//continue looking or continue skipping (if goOn is false)
[
{current: 0, currentPeriod: "School is out", remaining: 0},//default value
true//initial value for goOn
]
)[0];//reduced to multiple values (value, go on) only need value
}
};
//some tests
const date = new Date();
//partially apply with some bell times
const schoolTime = getInfoSchoolTime(
[
"8:10", "9:02", "9:54", "9:59", "10:51",
"11:43", "11:58", "12:48", "13:35", "13:40",
"14:10", "14:10", "15:02", "15:54"
]
);
//helper to log time from a date
const formatTime = date =>
("0"+date.getHours()).slice(-2)+":"+("0"+date.getMinutes()).slice(-2);
date.setHours(11);
date.setMinutes(1);
console.log(formatTime(date),schoolTime(date));//11:01
date.setHours(15);
date.setMinutes(53);
console.log(formatTime(date),schoolTime(date));//15:53
date.setHours(23);
date.setMinutes(1);
console.log(formatTime(date),schoolTime(date));//23:01
Consider this nested array of dates and names:
var fDates = [
['2015-02-03', 'name1'],
['2015-02-04', 'nameg'],
['2015-02-04', 'name5'],
['2015-02-05', 'nameh'],
['1929-03-12', 'name4'],
['2023-07-01', 'name7'],
['2015-02-07', 'name0'],
['2015-02-08', 'nameh'],
['2015-02-15', 'namex'],
['2015-02-09', 'namew'],
['1980-12-23', 'name2'],
['2015-02-12', 'namen'],
['2015-02-13', 'named'],
]
How can I identify those dates that are out of sequence. I don't care if dates repeat, or skip, I just need the ones out of order. Ie, I should get back:
results = [
['1929-03-12', 'name4'],
['2023-07-01', 'name7'],
['2015-02-15', 'namex'],
['1980-12-23', 'name2'],
]
('Namex' is less obvious, but it's not in the general order of the list.)
This appears to be a variation on the Longest Increase Subsequence (LIS) problem, with the caveat that there may be repeated dates in the sequence but shouldn't ever step backward.
Use case: I have sorted and dated records and need to find the ones where the dates are "suspicious" -- perhaps input error -- to flag for checking.
NB1: I am using straight Javascript and NOT a framework. (I am in node, but am looking for a package-free solution so I can understand what's going on...)
Here's an adaptation of Rosetta Code LIS to take a custom getElement and compare functions. We can refine the comparison and element-get functions based on your specific needs.
function f(arr, getElement, compare){
function findIndex(input){
var len = input.length;
var maxSeqEndingHere = new Array(len).fill(1)
for(var i=0; i<len; i++)
for(var j=i-1;j>=0;j--)
if(compare(getElement(input, i), getElement(input, j)) && maxSeqEndingHere[j] >= maxSeqEndingHere[i])
maxSeqEndingHere[i] = maxSeqEndingHere[j]+1;
return maxSeqEndingHere;
}
function findSequence(input, result){
var maxValue = Math.max.apply(null, result);
var maxIndex = result.indexOf(Math.max.apply(Math, result));
var output = new Set();
output.add(maxIndex);
for(var i = maxIndex ; i >= 0; i--){
if(maxValue==0)break;
if(compare(getElement(input, maxIndex), getElement(input, i)) && result[i] == maxValue-1){
output.add(i);
maxValue--;
}
}
return output;
}
var result = findIndex(arr);
var final = findSequence(arr, result)
return arr.filter((e, i) => !final.has(i));
}
var fDates = [
['2015-02-03', 'name1'],
['2015-02-04', 'nameg'],
['2015-02-04', 'name5'],
['2015-02-05', 'nameh'],
['1929-03-12', 'name4'],
['2023-07-01', 'name7'],
['2015-02-07', 'name0'],
['2015-02-08', 'nameh'],
['2015-02-15', 'namex'],
['2015-02-09', 'namew'],
['1980-12-23', 'name2'],
['2015-02-12', 'namen'],
['2015-02-13', 'named'],
];
console.log(f(fDates, (arr, i) => arr[i][0], (a,b) => a >= b));
This solution tries to get all valid sequences and returns the longes sequences for filtering the parts out.
It works by iterating the given array and checks if the values could build a sequence. If a value is given, which part result has a valid predecessor, the array is appended with this value. If not a backtracking is made and a sequence is searched with a valid predecessor.
act. array
value 7 3 4 4 5 1 23 7 comment
----- ------------------------ ---------------------------
7 7 add array with single value
3 7 keep
3 add array with single value
4 7 keep
3 4 add value to array
4 7 keep
3 4 4 add value to array
5 7 keep
3 4 4 5 add value to array
1 7 keep
3 4 4 5 keep
1 add array with single value
23 7 23 add value to array
3 4 4 5 23 add value to array
1 23 add value to array
7 7 23 keep
7 7 fork above, filter for smaller or equal and add value
3 4 4 5 23 keep
3 4 4 5 7 fork above, filter for smaller or equal and add value
1 23 keep
1 7 fork above, filter for smaller or equal and add value
function longestSequences(array, getValue = v => v) {
return array
.reduce(function (sub, value) {
var single = true;
sub.forEach(function (s) {
var temp;
if (getValue(s[s.length - 1]) <= getValue(value)) {
s.push(value);
single = false;
return;
}
// backtracking
temp = s.reduceRight(function (r, v) {
if (getValue(v) <= getValue(r[0])) {
r.unshift(v);
single = false;
}
return r;
}, [value]);
if (temp.length !== 1 && !sub.some(s => s.length === temp.length && s.every((v, i) => getValue(v) === getValue(temp[i])))) {
sub.push(temp);
}
});
if (single) {
sub.push([value]);
}
return sub;
}, [])
.reduce(function (r, a) {
if (!r || r[0].length < a.length) {
return [a];
}
if (r[0].length === a.length) {
r.push(a);
}
return r;
}, undefined);
}
function notInSequence(array, getValue = v => v) {
var longest = longestSequences(array, getValue);
return array.filter((i => a => a !== longest[0][i] || !++i)(0));
}
var array = [7, 3, 4, 4, 5, 1, 23, 7, 8, 15, 9, 2, 12, 13],
fDates = [['2015-02-03', 'name1'], ['2015-02-04', 'nameg'], ['2015-02-04', 'name5'], ['2015-02-05', 'nameh'], ['1929-03-12', 'name4'], ['2023-07-01', 'name7'], ['2015-02-07', 'name0'], ['2015-02-08', 'nameh'], ['2015-02-15', 'namex'], ['2015-02-09', 'namew'], ['1980-12-23', 'name2'], ['2015-02-12', 'namen'], ['2015-02-13', 'named']],
usuallyFailingButNotHere = [['2015-01-01'], ['2014-01-01'], ['2015-01-02'], ['2014-01-02'], ['2015-01-03'], ['2014-01-03'], ['2014-01-04'], ['2015-01-04'], ['2014-01-05'], ['2014-01-06'], ['2014-01-07'], ['2014-01-08'], ['2014-01-09'], ['2014-01-10'], ['2014-01-11']],
test2 = [['1975-01-01'], ['2015-02-03'], ['2015-02-04'], ['2015-02-04'], ['2015-02-05'], ['1929-03-12'], ['2023-07-01'], ['2015-02-07'], ['2015-02-08']];
console.log(longestSequences(array));
console.log(notInSequence(array));
console.log(notInSequence(fDates, a => a[0]));
console.log(longestSequences(usuallyFailingButNotHere, a => a[0]));
console.log(notInSequence(usuallyFailingButNotHere, a => a[0]));
console.log(longestSequences(test2, a => a[0]));
console.log(notInSequence(test2, a => a[0]));
.as-console-wrapper { max-height: 100% !important; top: 0; }
This solution uses the function reduce and keeps the previously accepted date to make the necessary comparisons.
var fDates = [['2015-02-03', 'name1'], ['2015-02-04', 'nameg'], ['2015-02-04', 'name5'], ['2015-02-05', 'nameh'], ['1929-03-12', 'name4'], ['2023-07-01', 'name7'], ['2015-02-07', 'name0'], ['2015-02-08', 'nameh'], ['2015-02-15', 'namex'], ['2015-02-09', 'namew'], ['1980-12-23', 'name2'], ['2015-02-12', 'namen'], ['2015-02-13', 'named']],
results = fDates.reduce((acc, c, i, arr) => {
/*
* This function finds a potential valid sequence.
* Basically, will check if any next valid sequence is
* ahead from the passed controlDate.
*/
function sequenceAhead(controlDate) {
for (var j = i + 1; j < arr.length; j++) {
let [dt] = arr[j];
//The controlDate is invalid because at least a forward date is in conflict with its sequence.
if (dt > acc.previous && dt < controlDate) return true;
}
//The controlDate is valid because forward dates don't conflict with its sequence.
return false;
}
let [date] = c; //Current date in this iteration.
if (i > 0) { // If this is not the first iteration
if (date === acc.previous) return acc; // Same as previous date are skipped.
// If the current date is lesser than previous then is out of sequence.
// Or if there is at least valid sequence ahead.
if (date < acc.previous || sequenceAhead(date)) acc.results.push(c);
else acc.previous = date; // Else, this current date is in sequence.
}
else acc.previous = date; // Else, set the first date.
return acc;
}, { 'results': [] }).results;
console.log(results);
.as-console-wrapper { max-height: 100% !important; top: 0; }
All of previous answers focus on JavaScript and maybe they won't work
correctly. So I decided to add new answer that focused on
Algorithm.
As #Trees4theForest mentioned in his question and comments, he is looking for a solution for Longest Increase Subsequence and out of order dates are dates that aren't in Longest Increase Subsequence (LIS) set.
I used this method like below. In algorithm's point of view, it's true.
function longestIncreasingSequence(arr, strict) {
var index = 0,
indexWalker,
longestIncreasingSequence,
i,
il,
j;
// start by putting a reference to the first entry of the array in the sequence
indexWalker = [index];
// Then walk through the array using the following methodolgy to find the index of the final term in the longestIncreasing and
// a sequence (which may need altering later) which probably, roughly increases towards it - http://en.wikipedia.org/wiki/Longest_increasing_subsequence#Efficient_algorithms
for (i = 1, il = arr.length; i < il; i++) {
if (arr[i] < arr[indexWalker[index]]) {
// if the value is smaller than the last value referenced in the walker put it in place of the first item larger than it in the walker
for (j = 0; j <= index; j++) {
// As well as being smaller than the stored value we must either
// - be checking against the first entry
// - not be in strict mode, so equality is ok
// - be larger than the previous entry
if (arr[i] < arr[indexWalker[j]] && (!strict || !j || arr[i] > arr[indexWalker[j - 1]])) {
indexWalker[j] = i;
break;
}
}
// If the value is greater than [or equal when not in strict mode) as the last in the walker append to the walker
} else if (arr[i] > arr[indexWalker[index]] || (arr[i] === arr[indexWalker[index]] && !strict)) {
indexWalker[++index] = i;
}
}
// Create an empty array to store the sequence and write the final term in the sequence to it
longestIncreasingSequence = new Array(index + 1);
longestIncreasingSequence[index] = arr[indexWalker[index]];
// Work backwards through the provisional indexes stored in indexWalker checking for consistency
for (i = index - 1; i >= 0; i--) {
// If the index stored is smaller than the last one it's valid to use its corresponding value in the sequence... so we do
if (indexWalker[i] < indexWalker[i + 1]) {
longestIncreasingSequence[i] = arr[indexWalker[i]];
// Otherwise we need to work backwards from the last entry in the sequence and find a value smaller than the last entry
// but bigger than the value at i (this must be possible because of the way we constructed the indexWalker array)
} else {
for (j = indexWalker[i + 1] - 1; j >= 0; j--) {
if ((strict && arr[j] > arr[indexWalker[i]] && arr[j] < arr[indexWalker[i + 1]]) ||
(!strict && arr[j] >= arr[indexWalker[i]] && arr[j] <= arr[indexWalker[i + 1]])) {
longestIncreasingSequence[i] = arr[j];
indexWalker[i] = j;
break;
}
}
}
}
return longestIncreasingSequence;
}
With method above, we can find dates that is out of order like below:
// Finding Longest Increase Subsequence (LIS) set
var _longestIncreasingSequence = longestIncreasingSequence(fDates.map(([date]) => date));
// Out of order dates
var result = fDates.filter(([date]) => !_longestIncreasingSequence.includes(date));
Online demo(jsFiddle)
here is a simple self- explanatory solution. hope it will help you.
const findOutOfSequenceDates = items => {
items = items.map(d => d);
const sequence = [], outOfsequence = [];
sequence.push(items.shift());
const last = ind => sequence[sequence.length - ind][0];
items.forEach(item => {
const current = new Date(item[0]);
if (current >= new Date(last(1))) {
sequence.push(item);
} else if (current >= new Date(last(2))) {
outOfsequence.push(sequence.pop());
sequence.push(item);
} else {
outOfsequence.push(item);
}
});
return outOfsequence;
};
var fDates = [
['2015-02-03', 'name1'],
['2015-02-04', 'nameg'],
['2015-02-04', 'name5'],
['2015-02-05', 'nameh'],
['1929-03-12', 'name4'],
['2023-07-01', 'name7'],
['2015-02-07', 'name0'],
['2015-02-08', 'nameh'],
['2015-02-15', 'namex'],
['2015-02-09', 'namew'],
['1980-12-23', 'name2'],
['2015-02-12', 'namen'],
['2015-02-13', 'named'],
];
console.log(findOutOfSequenceDates(fDates));
Use the Javascript Date type. Compare with those objects. Very simplistically,
date1 = new Date(fDates[i, 0])
date2 = new Date(fDates[i+1, 0])
if (date2 < date1) { // or whatever comparison you want ...
// flag / print / alert the date
}
To clarify, This merely finds items out of sequence. You can do that with strings, as Jaromanda X pointed out. However, you use the phrase "way out of line"; whatever this means for you, Date should give you the ability to determine and test for it. For instance, is '2023-07-01' unacceptable because it's 8 years away, or simply because it's out of order with the 2015 dates? You might want some comparison to a simpler time span, such as one month, where your comparison will looks something like
if (date2-date1 > one_month)
Summary of your question
If I have understood your question correctly, you are trying to identify array entries that do not follow a chronological order based on the time/date property value.
Solution
Convert the date string / time into a UNIX time stamp (number of seconds lapsed since 01/jan/1970 at 00:00:00)
Using a loop, we can store the value against a previous reading per itenary, if the value is negative, this would indicate an error in the date lapse, if the value is positive, it would indicate the order is valid
When negative, we can create an array to denote the position of the reference array and its values allowing you to go back to the original array and review the data.
Example Code
var arrData = [
{date: '2015-02-03', value:'name1'},
{date: '2015-02-04', value:'nameg'},
{date: '2015-02-04', value:'name5'},
{date: '2015-02-05', value:'nameh'},
{date: '1929-03-12', value:'name4'},
{date: '2023-07-01', value:'name7'},
{date: '2015-02-07', value:'name0'},
{date: '2015-02-08', value:'nameh'},
{date: '2015-02-15', value:'namex'},
{date: '2015-02-09', value:'namew'},
{date: '1980-12-23', value:'name2'},
{date: '2015-02-12', value:'namen'},
{date: '2015-02-13', value:'named'}
];
var arrSeqErrors = [];
function funTestDates(){
var intLastValue = 0, intUnixDate =0;
for (x = 0; x <= arrData.length-1; x++){
intUnixDate = Date.parse(arrData[x].date)/1000;
var intResult = intUnixDate - intLastValue;
if (intResult < 0){
console.log("initeneration: " + x + " is out of sequence");
arrSeqErrors.push (arrData[x]);
}
intLastValue = intResult;
}
console.log("Items out of sequence are:");
console.log(arrSeqErrors);
}
funTestDates();