Data is not adjusted for splits despite adjustment="split" flag

I am getting unadjusted values despite using the adjustment=split option.

df = alpaca_api.get_bars(
                ["RMBL"],
                TimeFrame(1, TimeFrameUnit.Day),
                "2020-05-15",
                "2020-06-30",
                adjustment="split",
            ).df

As you see the split happened on the 2020-05-20, but is not adjusted, which renders my ML model useless:

Screenshot 2021-12-19 at 09.45.52

Hey :wave:

This seems to have been resolved since, right?

I’m seeing this same issue today, on TQQQ, looking at the split which occurred on 2022-01-13. I’ve tried raw, split, dividend, and all.

Update: I was using the “SIP” data feed for my requests, but I don’t currently have a paid SIP subscription. Switching to the “IEX” feed made the “split” adjustment work as expected. After digging a little bit deeper, I realized that it’s because SIP has more data, including 00:00:00 - 00:59:59 on 1/13/2022. This one hour, for whatever reason, is unadjusted for the split.

I have the same issue. If it helps, I’ve created a script to identify stocks with a bug in the adjustment, and I’ve identified the following symbols. I hope this is especially useful for those applying ML models.

MISS_ADJ = [
“SIRI”, #2024-09-09
“GSAT”, #2021-02-08
“XPRO”, #2021-10-04
“TT”, #2020-03-02,
“EGO”, #2018-12-31
“BW”, #2019-07-22
“ELP”, #2021-04-26
“RCEL”, #2020-06-29,
“AIFU”, #2025-05-19,
“ALT”, #2018-09-17
“CIVI” #2017-05-01
]