Why does my daily median calculation and write to channel fail unpredictably?

I am trying to calculate the daily median across a range of channels and write the result to a new field for each channel. The code seems to run sometimes, but runs into rate limitation other times. I'm not sure why since I am only writing once and to different channels?

Error using Median calcs (line 39)

Requests are too frequent. For further information, see Limitations in the documentation.

My code is:

% Sensor/channel configuration
sensors = [
    struct('name', 'McKinley','channelID', 28xxxx, 'writeKey', 'G3WPJ586M55Gxxxx')
    struct('name', 'DPI', 'channelID', 80xxxx, 'writeKey', 'S0E0LB45GQLMxxxx')
    struct('name', 'Bryony','channelID', 29xxxx, 'writeKey', '2BPCI0IOAINPxxxx')
];

% Define date range (last 24 hours to midnight)
endTime = dateshift(datetime('now'), 'start', 'day');
startTime = endTime - hours(24);

% Preallocate a results array
results = [];

% Step 1: Read and calculate all medians
for i = 1:length(sensors)
    s = sensors(i);
    
    % Read Field 1 for the past 24 hours
    [weight, ~] = thingSpeakRead(s.channelID, 'DateRange', [startTime, endTime], 'Fields', 1);
    
    % Compute median, ignoring NaNs
    medianWeight = round(median(weight, 'omitnan'),2);
    
    % Store in results
    results(i).name = s.name;
    results(i).channelID = s.channelID;
    results(i).writeKey = s.writeKey;
    results(i).value = medianWeight;
end

% Step 2: Display results
for i = 1:length(results)
    r = results(i);
    fprintf('%s (Channel %d) → 24hr Median: %.2f\n', r.name, r.channelID, r.value);
end

% Step 3: Write results
    thingSpeakWrite(results(1).channelID, 'Fields', 6, 'Values', {results(1).value}, 'WriteKey', results(1).writeKey);
    thingSpeakWrite(results(2).channelID, 'Fields', 6, 'Values', {results(2).value}, 'WriteKey', results(2).writeKey);
    thingSpeakWrite(results(3).channelID, 'Fields', 6, 'Values', {results(3).value}, 'WriteKey', results(3).writeKey);

[Christopher Stapels]

Writing a table of values from MATLAB uses the bulk write interface.

See the limitations area there for additional info. How often are you running this code?

"The number of messages in a single bulk-update is limited to 960 messages for users of free accounts and 14,400 messages for users of paid accounts. The time interval between sequential bulk-update calls should be 15 seconds or more."

Answer 1

[Ray Berkelmans]

I'm running the code once per day, about a minute after midnight. So far it has run 6 out of 9 nights successfully and failed 3 nights. I'm not sure if there is something in the pattern, but so far it is 2 success, 1 fail, 2 success, 1 fail, 2 success, 1 fail. Is this co-incidental?

I have a paid account, so could easily add a 1 sec delay between writes. I'll try that and see how it goes.

[Vinod]

If you don't need to run it at exactly midnight, I'd also suggest some more random time (like 12:14 AM, for example). This will avoid the issue of potentially running into jitter from other work which is also scheduled exactly at midnight.

[Ray Berkelmans]

Yep, will do!

Answer 2

[Heidi]

A combination of 1s delay between writes and moving the TimeControl away from midnight seems to be doing the trick. Three nights good so far... :)