Price Repair#
The new argument repair=True in history() and download() will attempt to fix a variety of price errors caused by Yahoo. Only US market data appears perfect, I guess Yahoo doesn’t care much about rest of world?
The returned table will have a new column Repaired? that specifies if row was repaired.
Price repair#
Missing dividend adjustment#
If dividend in data but preceding Adj Close = Close, then manually apply dividend-adjustment to Adj Close.
Note: Repaired? is NOT set to True because fix only changes Adj Close
8TRA.DE#
Missing split adjustment#
If stock split in data but preceding price data is not adjusted, then manually apply stock split. Requires date range include 1 day after stock split for calibration - sometimes Yahoo fails to adjust prices on stock split day.
MOB.ST#
Missing data#
If price data is clearly missing or corrupt, then reconstructed using smaller interval e.g. 1h to fix 1d data.
1COV.DE missing row#
1COV.DE missing Volume, but intraday price changed#
0316.HK missing Volume, but daily price changed#
100x errors#
Sometimes Yahoo mixes up currencies e.g. $/cents or £/pence. So some prices are 100x wrong.
Sometimes they are spread randomly through data - these detected with scipy module.
Other times they are in a block, because Yahoo decided one day to permanently switch currency.
AET.L#
Price reconstruction - algorithm notes#
Spam minimised by grouping fetches. Tries to be aware of data limits e.g. 1h cannot be fetched beyond 2 years.
If Yahoo eventually does fix the bad data that required reconstruction, you will see it’s slightly different to reconstructed prices and volume often significantly different. Best I can do, and beats missing data.
Dividend repair (new)#
Fix errors in dividends:
adjustment missing or 100x too small/big for the dividend
duplicate dividend (within 7 days)
dividend 100x too big/small for the ex-dividend price drop
ex-div date wrong (price drop is few days/weeks after)
Most errors I’ve seen are on London stock exchange (£/pence mixup), but no exchange is safe.
IMPORTANT - false positives#
Because fixing (3) relies on price action, there is a chance of a “false positive” (FP) - thinking an error exists when data is good. FP rate increases with longer intervals, so only 1d intervals are repaired. If you request repair on multiday intervals (weekly etc), then: 1d is fetched from Yahoo, repaired, then resampled - this has nice side-effect of solving Yahoo’s flawed way of div-adjusting multiday intervals.
FP rate on 1d is tiny. They tend to happen with tiny dividends e.g. 0.5%, mistaking normal price volatility for an ex-div drop 100x bigger than the dividend, causing repair of the “too small” dividend (repair logic already tries to account for normal volatility by subtracting median). Either accept the risk, or fetch 6-12 months of prices with at least 2 dividends - then can analyse the dividends together to identify false positives.
Adjustment missing#
1398.HK
# ORIGINAL:
Close Adj Close Dividends
2024-07-08 00:00:00+08:00 4.33 4.33 0.335715
2024-07-04 00:00:00+08:00 4.83 4.83 0.000000
# REPAIRED:
Close Adj Close Dividends
2024-07-08 00:00:00+08:00 4.33 4.330000 0.335715
2024-07-04 00:00:00+08:00 4.83 4.494285 0.000000
Adjustment too small#
3IN.L
# ORIGINAL:
Close Adj Close Dividends
2024-06-13 00:00:00+01:00 3.185 3.185000 0.05950
2024-06-12 00:00:00+01:00 3.270 3.269405 0.00000
# REPAIRED:
Close Adj Close Dividends
2024-06-13 00:00:00+01:00 3.185 3.185000 0.05950
2024-06-12 00:00:00+01:00 3.270 3.210500 0.00000
Duplicate (within 7 days)#
ALC.SW
# ORIGINAL:
Close Adj Close Dividends
2023-05-10 00:00:00+02:00 70.580002 70.352142 0.21
2023-05-09 00:00:00+02:00 65.739998 65.318443 0.21
2023-05-08 00:00:00+02:00 66.379997 65.745682 0.00
# REPAIRED:
Close Adj Close Dividends
2023-05-10 00:00:00+02:00 70.580002 70.352142 0.00
2023-05-09 00:00:00+02:00 65.739998 65.527764 0.21
2023-05-08 00:00:00+02:00 66.379997 65.956371 0.00
Dividend too big#
HLCL.L
# ORIGINAL:
Close Adj Close Dividends
2024-06-27 00:00:00+01:00 2.360 2.3600 1.78
2024-06-26 00:00:00+01:00 2.375 2.3572 0.00
# REPAIRED:
Close Adj Close Dividends
2024-06-27 00:00:00+01:00 2.360 2.3600 0.0178
2024-06-26 00:00:00+01:00 2.375 2.3572 0.0000
Dividend & adjust too big#
LTI.L
# ORIGINAL:
Close Adj Close Adj Dividends
2024-08-08 00:00:00+01:00 768.0 768.0 1.0000 5150.0
2024-08-07 00:00:00+01:00 819.0 -4331.0 -5.2882 0.0
Close Adj Close Adj Dividends
2024-08-08 00:00:00+01:00 768.0 768.0 1.0000 51.5
2024-08-07 00:00:00+01:00 819.0 767.5 0.9371 0.0
Dividend too small#
BVT.L
# ORIGINAL:
Close Adj Close Adj Dividends
2022-02-03 00:00:00+00:00 0.7534 0.675197 0.8962 0.00001
2022-02-01 00:00:00+00:00 0.7844 0.702970 0.8962 0.00000
# REPAIRED:
Close Adj Close Adj Dividends
2022-02-03 00:00:00+00:00 0.7534 0.675197 0.8962 0.001
2022-02-01 00:00:00+00:00 0.7844 0.702075 0.8950 0.000
Adjusted 2x on day before#
clue: Close < Low
2020.OL
# ORIGINAL:
Low Close Adj Close Dividends
2023-12-21 00:00:00+01:00 120.199997 121.099998 118.868782 0.18
2023-12-20 00:00:00+01:00 122.000000 121.900002 119.477371 0.00
# REPAIRED:
Low Close Adj Close Dividends
2023-12-21 00:00:00+01:00 120.199997 121.099998 118.868782 0.18
2023-12-20 00:00:00+01:00 122.000000 122.080002 119.654045 0.00
ex-div date wrong#
TETY.ST
# ORIGINAL:
Close Adj Close Dividends
2022-06-22 00:00:00+02:00 66.699997 60.085415 0.0
2022-06-21 00:00:00+02:00 71.599998 64.499489 0.0
2022-06-20 00:00:00+02:00 71.800003 64.679657 5.0
2022-06-17 00:00:00+02:00 71.000000 59.454838 0.0
# REPAIRED:
Close Adj Close Dividends
2022-06-22 00:00:00+02:00 66.699997 60.085415 5.0
2022-06-21 00:00:00+02:00 71.599998 60.007881 0.0
2022-06-20 00:00:00+02:00 71.800003 60.175503 0.0
2022-06-17 00:00:00+02:00 71.000000 59.505021 0.0