Datasets:
The data shape for the most recent 3 months of data is different to the previous months
The 30-minutely data for 2024-12, 2025-01, and 2025-02 looks like this:
date,ss_id,t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12,t13,t14,t15,t16,t17,t18,t19,t20,t21,t22,t23,t24,t25,t26,t27,t28,t29,t30,t31,t32,t33,t34,t35,t36,t37,t38,t39,t40,t41,t42,t43,t44,t45,t46,t47,t48,datetime_GMT
2024-12-02 00:00:00,2405,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,9.25,6.33,104.94,90.58,51.63,55.45,57.32,121.51,38.44,25.45,4.95,6.3,8.49,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,2024-12-02T00:00:00.000000000+0000
2024-12-02 00:00:00,2428,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,6.92,31.78,61.46,135.28,71.49,71.83,62.67,58.62,50.57,45.93,22.66,30.69,6.41,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,2024-12-02T00:00:00.000000000+0000
2024-12-02 00:00:00,2429,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.04,14.76,14.66,16.36,24.2,40.91,25.29,26.5,24.72,45.6,18.35,13.69,6.16,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,2024-12-02T00:00:00.000000000+0000
2024-12-02 00:00:00,2430,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.99,6.99,14.71,19.14,25.13,42.66,25.72,28.91,25.68,55.64,16.43,16.59,4.47,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,2024-12-02T00:00:00.000000000+0000
2024-12-02 00:00:00,2431,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,22.1,11.87,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,2024-12-02T00:00:00.000000000+0000
In contrast, the data for previous months looks like this:
| ss_id (i64) | datetime_GMT | generation_Wh (f64) |
|---|---|---|
| 2405 | 2018-01-01T00:30:00Z | 0.0 |
| 2406 | 2018-01-01T00:30:00Z | 0.0 |
| 2408 | 2018-01-01T00:30:00Z | 0.0 |
| 2409 | 2018-01-01T00:30:00Z | 0.0 |
| 2410 | 2018-01-01T00:30:00Z | 0.0 |
Questions:
- Does each column
t1,t2, ...,t48represent the watt hours generated for each half-hour period of the day? If so, doest1represent the total energy generated from00:00to00:30? - Is this how all the data will look, going forwards?
Thanks for the heads up Jack - I'm presuming your assumption 1 is correct, but as for 2, that's up for discussion! Doesn't feel like it would be too hard to move back to the old structure, for ease of integration?
Awesome, thanks for the very quick reply, @sol-ocf !
Thanks for confirming that this change happened upstream of OCF.
Yeah, I agree, it'd be nice to keep everything in the same format, to make our users' lives as easy as possible. I'm happy to write some code (in Polars, if that's OK) to convert the new data into the old structure. But, before I do that, we should probably decide on whether we're happy to pivot all the timeseries data so we store the data as very wide tables (as proposed in issue #14).
I've written some Polars code to convert Sheffield Solar's new "wide" format (where each half hour is a different column) into the "tall and thin" format that was used in previous months. Once I've checked that the newly converted data passes some sanity checks, I'll push it to the repo.
Note that Sheffield Solar's new "wide" format contains quite a lot of duplicates. My code de-duplicates the data.
OK, I've pushed the reshaped data for the last 3 months in git commit 0c1e859c7e0a3f8c167209eefd69efee65590be2