Read Raw Historical Data

  • Can return Historical data points or a query strategy.

  • Returns by default a query strategy when the total number of data points exceeds three million.

  • Can be forced to return less data points by setting the 'query_count_limit'

  • Can be used to fetch historical data in reversed order by switching the start and end time.

  • Accepts a filter for server side filtering.

  • Accepts 'Match Duration' filters to perform server side duration calculations.

  • Required security permission - READ.

  • HTTP methods supported: GET, POST.

Description

This endpoint allows to read raw historical data for provided time period for one or more items.

Parameters

Name Type Optional Description

start_time

string

no

The start time (UTC) of the interval to retrieve data for.

end_time

string

no

The end time (UTC) of the interval to retrieve data for.

identifier

string

no

The path or ID of the object to be queried for historical information.

filter

string

yes

Filter query. Has to be a Base64 encoded JSON (dictionary).

fields

string

yes

Comma separated string to customize which fields should be included in the response, by default 'i,p,v,q,t'. Provide 'ALL' to include all fields.

bounds

undefined

yes

A Boolean or number, which indicates bounding values should be returned, by default false. 0 / false = None, 1 / true = both start and end, 2 = only start, 3 = only end.

query_count_limit

integer

yes

The maximum number of data points per query. Optional, by default 3000000.

query_index

string

yes

The index of the query to retrieve historical data for. In case a query index is provided, the query will be executed directly, without determining a strategy first.

insights

boolean

yes

Indicates application insights should be included in the response, by default false.

Filtering

The filter option can either be an Aggregation Pipeline with the supported stages '\($match' and '$\)limit' or a match query. The syntax correspond respectively to the MongoDB’s Aggregation Pipeline or read operation query syntax.

Supported Pipeline Stages

Stage Description

$match

Filters the historical data points.

$limit

Limits the number of matching historical data points. The first n data points will be returned, where n is the specified limit.

Supported query selectors are:

Comparison

Name Description

$eq

Matches values that are equal to a specified value.

$gt

Matches values that are greater than a specified value.

$gte

Matches values that are greater than or equal to a specified value.

$in

Matches any of the values specified in an array.

$lt

Matches values that are less than a specified value.

$lte

Matches values that are less than or equal to a specified value.

$ne

Matches all values that are not equal to a specified value.

$nin

Matches none of the values specified in an array.

Logical

Name Description

$and

Joins query clauses with a logical AND returns all documents that match the conditions of both clauses.

$or

Joins query clauses with a logical OR returns all documents that match the conditions of either clause.

Query Strategy

By default read raw historical data returns a query strategy when the total number of data points exceeds three million. If "query_count_limit" is set to lower number of data points than requested data point number, response will also be a query strategy. To get the usual "v", "q", "t" response structure read raw historical data endpoint can be called providing the query index as a parameter for GET call or the query as JSON dictionary as a request body for POST call.

Duration calculations

With the field option 'd' it is possible to request the server to calculate the duration between each historical data point. The first duration is calculated from the interval start time, the last duration is calculated till the end time of the requested interval.

Interval                         Interval
 Start                             End
   +                                +
40 |            x                   |
30 |                          x     |
20 |                   x            |
10 |     x                          |
   +---------------------------------

   <----> <----> <----> <----> <---->
    d[1]   d[2]   d[3]   d[4]   d[5]

Match Duration

The 'Match Duration' feature gives the possibility to perform (server-side) duration calculations based on a match expression. To be able to calculate the duration for the first and last data point in the requested interval the first matching bounding values will automatically be fetched based on the provided Match duration options:

Name Description

leading_bounding_time_span

The maximum time span in milliseconds before the leading bounding value of the requested interval to search for a matching bounding value.

leading_bounding_intervals_no

The number of intervals to use to search for a matching leading bounding value.

trailing_bounding_time_span

The maximum time span in milliseconds after the trailing bounding value of the requested interval to search for a matching bounding value.

trailing_bounding_intervals_no

The number of intervals to use to search for a trailing leading bounding value.

The following example shows how Match Durations can be defined in a 'ReadRawHistoricalDataQuery' / request body:

{
    "md": [
        {
            "name": "MyDuration01",
            "expression": {
                "v": {
                    "$gt": 20
                }
            },
            "leading_bounding_time_span": 15000,
            "trailing_bounding_time_span": 15000,
            "leading_bounding_intervals_no": 3,
            "trailing_bounding_intervals_no": 3
        }
    ]
}

For example, setting the 'leading_bounding_time_span' 15000 and the 'leading_bounding_intervals_no' to 3, means the logic searches for a matching leading bounding value in three steps starting from the timestamp of the leading bounding value of the requested interval.

     matching
     leading                              interval
     bounding                    leading   start
      value                      bounding    +
        x                         value      |
                                    .<------>+
                                      3000ms |
                                             |
      <------------------------------------->+---+
          leading bounding time span

      <----><----------><---------->
     2000ms  5000ms      5000ms
(interval 3)(interval 2)(interval 1)

Data point 'x' is included in the response and used in the match duration calculation. Data point '.' is excluded from the response and match duration calculation.

Examples

The following example shows the output based on the Match Duration expression '{"v" : {"$gt": 20}}':

        interval                         interval
         start                             end
           +                                +
x       40 |            x                   |        x
        30 |                          x     |
     *  20 |                   *            |  *
        10 |     *                          |
           +--------------------------------+

 <---------------------> <-----------> <------------>
  d[1]                    d[2]          d[3]

Data points indicated by ' * ' are included in the response but excluded from the match duration calculation.

Match Duration calculations can be used together with the filter option. The following example shows the output based on the Match Duration expression '{"v" : {"\($gt": 20}}' and filter '{"v" : {"$\)gte": 20}}'

        interval                         interval
         start                             end
           +                                +
x       40 |            x                   |        x
        30 |                          x     |
     *  20 |                   *            |  *
        10 |     .                          |
           +--------------------------------+

 <---------------------> <-----------> <------------>
  d[1]                    d[2]          d[3]

Post Execution

The Post execution feature makes it possible to execute any Lua logic server-side (in the Core) to process / transform the data returned by a 'ReadRawHistoricalData' query.

The following example shows how Post Execution functions can be defined in a ReadRawHistoricalDataQuery / request body:

{
    "post_execution": [
    {
        "lib": "CalcLib",
        "func": "countOccurrences",
        "arg": {}
    }
  ]
}

The library defined in a Post Execution item must be available in the context of the Web API. An example library to count the number of historical data points per item can be found here

The implementation of the library

local lib = {}

function lib:countOccurrences(arg)
    arg = arg or {}
    local result = {}
    local data = arg.data or {}
    for _, item in ipairs(data.items) do
        local count = 0
        -- v can contain nil values, use t.
        if type(item.t) == 'table' then
            count = #item.t
        end
        table.insert(result, {
            p = item.p,
            v = count
        })
    end
    return result
end

return lib

Function argument

The raw historical data query result gets passed to the function as part of the arg argument. The JSON representation is as follows:

{
    "data": {
        "query_index": "1",
        "start_time": "2018-11-11T00:00:00.000Z",
        "end_time": "2018-11-13T00:00:00.000Z",
        "items": [
            {
            "p": "/System/Core/...",
            "v": [],
            "q": [],
            "t": []
            }
        ]
    }
}

Response content type

Post Execution makes it also possible to return the data in a different format then JSON. The example library rrhd-post-exec-with-headers show how to return a csv by defining the HTTP headers '["Content-Type"] = "txt/csv"' in the response.

Function chain

By defining multiple 'execFunction' items in the 'post_execution' array. A chain of functions can be created. During execution the result of a function will be passed to the next function in the data field as part the arg argument.