Downloading Activities to CSV header data change

I know this isn't the answer you're looking for, but I think part of the problem here is that the Connect CSV export only exports a limited number of activities, right? So even if you wanted to re-export *all* activities every time, you couldn't (afaik).

There's a 3rd party site called https://www.runalyze.com which syncs with your Connect account and allows you to export *all* activity (summary data) to CSV (it only takes a few seconds for several years worth of activities, for me).

If you're curious, this is the header for the CSV it generates (Profile > Export data > Basic data > Activities:

id,hash,time,timezoneOffset,created,edited,isPublic,isTrack,distance,s,elapsedTime,elevationUp,elevationDown,elevationUpFile,elevationDownFile,altitudeSource,climbScore,percentageHilly,uphillEfficiency,downhillEfficiency,vam,kcal,pulseAvg,pulseMax,recoveryHr,maxHeartRateDropFrom,maxHeartRateDrop,vo2max,vo2maxByTime,vo2maxWithElevation,useVO2max,fitVO2maxEstimate,fitRecoveryTime,fitTrimp,fitHrvAnalysis,fitTrainingEffect,fitAnaerobicTrainingEffect,fitPerformanceCondition,fitPerformanceConditionEnd,fitFtp,fitSweatLoss,fitDeveloperSessionFields,rpe,subjectiveFeeling,trimp,cadence,gap,xPace,xGap,aerobicDecouplingPace,aerobicDecouplingPower,power,powerWithZero,airPower,isPowerCalculated,xPower,xPowerWithZero,wPrimeBalancePowerAvg,wPrimeBalancePowerMin,wPrimeBalancePaceAvg,wPrimeBalancePaceMin,staminaAvg,staminaMin,staminaPotentialAvg,staminaPotentialMin,zoneDistributionPace,zoneDistributionHr,zoneDistributionPower,totalStrokes,swolf,wheelSize,poolLength,strideLength,groundcontact,groundcontactBalance,verticalOscillation,verticalRatio,avgImpactGsLeft,avgImpactGsRight,avgBrakingGsLeft,avgBrakingGsRight,avgFootstrikeTypeLeft,avgFootstrikeTypeRight,avgPronationExcursionLeft,avgPronationExcursionRight,avgLegSpringStiffness,avgFormPower,avgRespiratoryRate,avgLeftRightBalance,avgLeftTorqueEffectiveness,avgRightTorqueEffectiveness,avgLeftPedalSmoothness,avgRightPedalSmoothness,avgPlatformCenterOffsetLeft,avgPlatformCenterOffsetRight,avgLeftPowerPhaseStartAngle,avgLeftPowerPhaseEndAngle,avgLeftPowerPhasePeakStartAngle,avgLeftPowerPhasePeakEndAngle,avgRightPowerPhaseStartAngle,avgRightPowerPhaseEndAngle,avgRightPowerPhasePeakStartAngle,avgRightPowerPhasePeakEndAngle,cyclingTotalGrit,cyclingTotalGrit2,cyclingAvgFlow,jumps,peakEpoc,peakEpocEstimate,temperature,windSpeed,windDeg,humidity,cloudCover,pressure,uvIndex,ozone,isNight,weatherid,weatherSource,routeName,hasTrackdata,isVirtual,splits,splitsCustom,title,partner,notes,creator,creatorDetails,deviceId,source,externalId,lock,sportid,typeid,accountid,hrv_id,original_file,hrv_avgRR,hrv_rmssd,hrv_sdsd,hrv_sdnn,hrv_sdnn5min,hrv_pnn50,hrv_pnn20,hrv_percentageAnomalies,hrv_dfaA1Avg,hrv_dfaA1Min,hrv_dfaA1Max,hrv_dfaA1PAboveAet,hrv_dfaA1PBelowAnt,hrv_dfaA1PValid,hrv_dfaA1PArtifacts,hrv_estimatedAetPower,hrv_estimatedAetPowerFeasible,hrv_estimatedAetPace,hrv_estimatedAetPaceFeasible,hrv_estimatedAetHr,hrv_estimatedAetHrFeasible,race_officialDistance,race_officialTime,race_officialAscent,race_officialDescent,race_officiallyMeasured,race_name,race_placeTotal,race_placeGender,race_placeAgeclass,race_participantsTotal,race_participantsGender,race_participantsAgeclass,race_priorityLevel,race_status,race_accountid,race_activity_id,race_race_event_id,tags,climbs,equipmentIds

zoepup said:

I am hoping for a better answer.

Not sure whether my answer meets your expectations, but I do not think that the header is fixed, at all. It seems to be wildly depending not only on the type you select - general Cycling, MTB, Indoor Cycling, Road Cycling, etc, but especially on the data in individual activities. I am getting very different headers depending on the filters I use. Some examples below. The best solution would be assigning the columns dynamically by name

Cycling              MTB                 Road Cycling
--------------------------------------------------------
Activity Type        Activity Type       Activity Type
Date                 Date                Date
Favorite             Favorite            Favorite
Title                Title               Title
Distance             Distance            Distance
Calories             Calories            Calories
Time                 Time                Time
Avg HR               Avg HR              Avg Speed
Max HR               Max HR              Max Speed
Aerobic TE           Aerobic TE          Training Stress Score
Avg Speed            Avg Speed           Total Strokes
Max Speed            Max Speed           Decompression
Total Ascent         Total Ascent        Best Lap Time
Total Descent        Total Descent       Number of Laps
Avg Bike Cadence     Avg GAP             Moving Time
Max Bike Cadence     Training Stress Sc  Elapsed Time
Normalized Power     Grit
Training Stress Sc   Flow
Max Avg Power 20min  Min Temp
Avg Power            Decompression
Max Power            Best Lap Time
Grit                 Number of Laps
Flow Max             Temp
Total Strokes        Moving Time
Min Temp             Elapsed Time
Decompression        Min Elevation
Best Lap Time        Max Elevation
Number of Laps
Max Temp
Moving Time
Elapsed Time
Min Elevation
Max Elevation

The time and decimal format is more problem of Excel which wants to be smarter than it is able to be. It is much better when using Open Calc

Be that as it may, Garmin should be aware that most ppl will use Excel to look at CSV files, and to make sure their output works consistently in excel.

An easy way to do this would be to make sure all times are either in hh:mm:ss or hh:mm:ss.ss format.

In the example above: 0:04:32 would be unambiguously interpreted as 4 minutes and 32 seconds, and it would work fine when using math with another value like either 4:21.5 or 0:04:21.5.

FlowState said:

I think part of the problem here is that the Connect CSV export only exports a limited number of activities, right⁇

Well, you can export all activities if you scroll to the very end of the Activities list before hitting the Export CSV link.

If that's the case, it seems that would be the solution here.

Btw, I just tried apache openoffice on a CSV file containing splits exported from Connect.

If I don't enable the "detect special numbers" option, indeed it doesn't try to be smart, as you said. But in that case, the times and paces are just interpreted as strings, which means they're useless for my purposes: e.g. using functions such as AVERAGE or SUM to analyze the data.

If I do enable the "detect special numbers" option, it works just like Excel. i.e. something like 4:30, representing a split of 4 minutes and 30 seconds, will be interpreted as hh:mm, while 4:30.5 (4 minutes and 30.5 seconds) will be correctly interpreted as mm:ss.ss. Again, this makes it impossible to do math on the splits without manually editing the times to add a leading "0:" where necessary.

So I will reiterate my assertion that Garmin's CSV output for exported splits is inconsistent, incorrect, and not as useful as it could be.

It would be so easy for Garmin to normalize the output so that all times and paces are formatted as either hh:mm:ss or [hh:]mm:ss.ss but they aren't aware of the problem and they probably wouldn't care either way. Rn the CSV splits are ok as human-readable output, but less than useful for manipulation via software. But the whole point of the CSV export is to manipulate the data with software, right?

EDIT: looking at my activity history in Connect now, it looks like lap times are always displayed and exported with fractional seconds (e.g. 4:30.0), at least when they're not collapsed into intervals, so maybe this issue has been addressed after all (as mm:ss.ss is unambiguous). If so, thanks Garmin!

I had the same thought but I always export to CSV from the header screen   I don’t select road cycling, mtb, etc. it has always worked up until now. I realize I can s roll to end  and export everything again. It’s only 18 years of data. I

i will try exporting after filtering only road cycling and report back. 

f that fails it is probably easier to reorganize my old data to match the new format … until it changes again 

It changes whenever there is a new activity in the filtered visible results that contains data that were not previously available in the older activities. For example when using a new type of activity, new sensor, new watch or head unit, or when Garmin adds a new metrics.