Query to get time count from mongodb - php

Question

I have my data in mongodb database as
{
"_id": ObjectId("5319acf0b06f6e98371ca505"),
"field1": "",
"date": "07-03-2014",
"os": "android",
"time": "11:26:40",
"userid": "xxxx"
}
I wanted to get count of all records having time from 00 to 23.
I have written my query as
$time_start = date('00');
$time_end = date('23');
$keys = array("time" => 1);
$initial = array("counttime" => array());
$reduce = "function(obj, prev) {
if (obj.time != null)
if (obj.time instanceof Array) prev.counttime += obj.time.length;
else prev.counttime++; }";
$condition = array(
'condition' => array(
"time" => array(
'$gt' => $time_start,
'$lte' => $time_end
)
)
);
$g = $collection->group($keys, $initial, $reduce , $condition);`
I have tried to get hour but thats not working for me.
I have used as
$condition = array(
'condition' => array(
"HOUR(time)" => array(
'$gt' => $time_start,'$lte' => $time_end)
)
);
Anyone who can help me out?

Answer 1

You can do this using .aggregate(), which you really should favor over .group() in any case, as it uses native code rather than JavaScript for it's results:
db.collection.aggregate([
// Match one date (tell you later)
{ "$match": { "date": "07-03-2014" } },
// Project the hour
{ "$project": {
"hour": { "$substr": [ "$time", 0, 2 ] }
}},
// Group on the hour
{ "$group": {
"_id": "$hour",
"count": { "$sum": 1 }
}}
])
Now this relies on that you have your "time" stored as a string. So you $project the "hour" part of that "time" and then $group on it, keeping a count of records found.
I restricted this to one "date" because as with "time" this is also a string. Depending on what you want to do you can possibly get unexpected results from this, so generally speaking, using "strings" for date representation is not a good idea.
If you just had a proper BSON date in a field called "timestamp" then you can do things like this, where you can even set ranges:
var start_date = new Date("2014-02-01");
var end_date = new Date("2014-03-01");
db.collection.aggregate([
// Match on the date range
{ "$match": { "timestamp": { "$gte": start_date, "$lt": end_date } },
// Project day and hour
{ "$project": {
"day": {
"year": { "$year": "$timestamp" },
"month": { "$month": "$timestamp" },
"day": { "$dayOfMonth": "$timestamp" },
"hour": { "$hour": "$timestamp" }
}
}},
// Group on the day
{ "$group": {
"_id": "$day",
"count": { "$sum": 1 }
}}
])
But with your present date format, ranges are going to be a problem. So best to convert where you can.