**Most important metric: `close` (daily closing price).** I chose it because it is the main price series for judging trends since 2012, including the effect of the AI surge. `open`, `high` and `low` are close variants of it, and `volume` measures trading activity, not price. This is my judgment of what matters most, not something the data computes. **Highest closing prices** (top 5 rendered above) - The record is **SK Hynix (000660.KS), close 300,000 on 2025-07-14**. - The next four are also SK Hynix, all in July 2025: 298,500 (07-15), 297,000 (07-10), 296,000 (07-16) and 294,500 (07-11). - This fits the AI-driven surge in memory chips. The record highs are all within the last few weeks of the data, which ends 2025-08-04. **Lowest closing prices** (bottom 5 rendered above) - The record low is **BrainChip (BRN.AX), close 0.009265 on 2014-12-05**. - The next four (2014-12-08 through 2014-12-11) have the same close of 0.009265, so the low is a tie. - In those four rows `open`, `high`, `low` and `close` are identical and volume is 0. The price never moved, so these look like stale carried-forward quotes from an illiquid or barely traded stock, not real trading. **Caveats** - Prices appear to be in each exchange's local currency: SK Hynix in Korean won and BrainChip in Australian dollars. The raw maximum and minimum are therefore not comparable across companies, and the ~32-million-fold gap between them mostly reflects currency and price scale. Comparing companies properly needs returns or normalised prices, not raw levels. - The low end is dominated by zero-volume days. Excluding days with no volume would give a more meaningful minimum.