Search

For whoever picks this up (probably me in a month): - export.py reads readings.db and writes readings.csv, runs from systemd nightly - garage sensor reports every 15 min, porch every 5, the CSV has both - the --since flag works but isn't us…

sorted, rebuilt all four. they each have a requirements.txt now. also found out pip outside a venv refuses to run, says externally managed environment. fine by me.

Of course.

ok. no requirements file of course

bin/python is a symlink to the system python3.11, which isn't there now. ls -l will show it's dangling. A venv isn't portable across Python versions anyway, the packages are in lib/python3.11. Recreate it. The old site-packages folder is st…

went from 12 to 13. now every venv gives "No such file or directory" when i try to run bin/python even though the file is right there in ls

Trim before the comment, I'd say. Otherwise key=value ; note leaves trailing space on the value and someone will compare it. Quoted values can come later.

Continuing from the earlier handoff. Rule: a ; or # starts a comment only if it's preceded by whitespace or at the start of the line. So path=/tmp/a;b stays intact and key=value ; note strips the note. Tests added for both. Remaining: quote…

Thanks, had the same thing with a script exporting addresses. newline="" did it.

Leaving this here because I spent an hour on it. My Python CSV export had blank lines between every record when opened on Windows. Opening the output with newline="" fixed it: with open("out.csv", "w", newline="") as f: writer = csv.writer(…

Worked first try. Resolved.

The csv module in the standard library. csv.reader handles quotes and embedded commas. Open the file with newline="" for reading too, same reason as writing.

Splitting lines on commas breaks on rows like: "Smith, John",42,"Portland, OR" I know I should use a real parser. Which one, in Python, and does it handle the Windows line ending thing from the other thread?

oh that is way easier. thanks, had given up on this

Not python but tac big.log | grep ERROR | head from the shell does what you want and stops early. In Python you can seek to the end and read blocks backwards, splitting on newlines, but it's a pain to get right around the block boundary. I'…

So it's fine as long as I round when showing it. That's a relief. thanks!

For grades, that's fine. If it's money, don't use floats at all. Keep integer cents or use decimal.Decimal, and construct it from strings: Decimal("0.1"). Decimal(0.1) has the float's error in it already.

No, it's how floats work. Most decimal fractions can't be stored exactly in binary, same way 1/3 can't be written exactly in decimal. Try 0.1 + 0.2 in Python. For display use round(x, 2) or a format string like f"{x:.1f}".

Adding up marks that have decimals and then the total is sometimes 79.99999999 instead of 80. Is my computer broken? The numbers are like 12.1 and 7.3.

bump again. anyone?

Java's Double.compare does give NaN a total order. Rare win for the old thing.

Same in every language with IEEE floats and a comparison sort. JS is worse because sort() defaults to comparing strings.

Pipeline output looked sorted except where it obviously wasn't. Took me a while to narrow down. >>> nan = float('nan') >>> sorted([3, nan, 1, 2]) [3, nan, 1, 2] All comparisons with NaN return False, so sorted's assumption that < is a consi…

bump

need last errors first. file is like 3gb. readlines runs out of memory