Partially scrambled half hour data- Got an EV

You’re not alone in struggling with the timings of half-hourly (Hh) data. The problem is that the electricity industry works always on UTC (GMT), while people in real life work on local time, currently BST (GMT+1). Depending on where the data come from, they may or may not have been adjusted for the difference. 

The OVO ‘Day’ Usage page is adjusted, so the times shown are local (BST). This leads to strange things at the start and end of BST, as you’ll probably see if you look up the page for  26 March 2023 (I say ‘probably,’ because I’m not certain that every smart meter is configured in the same way as far as this is concerned). 

I too try to track my Hh figures, but I find it easier to piggyback on one of the third-party data aggregators to get them. I use n3rgy, by way of Guy Lipman’s excellent utilities. Here’s what the CSV for the same timezone change look like, C&P from Excel:

It’s probably best to be consistent by sticking to UTC everywhere. If there are any gaps in the DCC data (e.g. during a power cut), these are flagged on Lipman’s Admin page.

I can’t explain why you’re seeing displacements by a whole day, and can only suspect that something went wrong when transferring figures from your source to the worksheet. There is, of course, the confusion at the OVO account site where meter readings are variously timed at the midnight (UTC) at the start or the end of the day they’re labelled with.
  

As regards @Jeffus’ remark about whether there’s a difference between a day’s consumption as measured by differencing meter readings on the one hand and summing Hh data on the other, I can only say that the last couple of weeks’ data for my new meter differ at most by 1Wh on any day, and by 1Wh for all the days. This is clearly not significant. There is, however, a tiny but significant difference between the peak and offpeak totals for each day, with the peak totals over-reported (and offpeak under-reported) by an average of about 3Wh each day. I give the supplier the benefit of the doubt here, guessing that the difference arises from the difference in reporting times of the Hh data on the one hand and the register readings on the other. If I’ve done my sums properly, 3Wh corresponds to about 4 minutes by which the offpeak period is under-reported, i.e. about 0.9%. It’s neither here nor there for a customer like me, but it all adds up - in OVO’s favour - if this scenario is repeated for every E7 customer. I wonder if settlement depends on the same data timings.