rozzin has left #qi-hardware ["Stream closed by us: Replaced by new connection (conflict)"]
rozzin has joined #qi-hardware
infobot has quit [Read error: Connection reset by peer]
infobot has joined #qi-hardware
eightdot has quit [Ping timeout: 240 seconds]
eightdot has joined #qi-hardware
incomprehensibly has quit [Ping timeout: 250 seconds]
qi-bot has quit [Ping timeout: 276 seconds]
qi-bot has joined #qi-hardware
eintopf has quit [Ping timeout: 260 seconds]
eintopf has joined #qi-hardware
incomprehensibly has joined #qi-hardware
ecloud is now known as ecloud_wfh
jwhitmore has joined #qi-hardware
<xentrac>
kyak: "look" is a database index lookup tool
<xentrac>
look traditionally does a binary search on file offsets, so you can apply it to files much larger than memory. GNU look doesn't do that by default but does have an option to do so
<xentrac>
between look, sort, join, comm, and potentially awk to rearrange fields if join -o isn't sufficiently convenient, you can do quite a bit of ad-hoc relational database stuff in the Unix shell
<xentrac>
even without the various demon spawn of dbm
<kyak>
so, stuff you can do in sqlite nowadays :)
<xentrac>
yes, and you probably should in many cases
<kyak>
but it's interesting, still
<xentrac>
but you don't get a query optimizer, you don't get automatic index updates (make is not really adequate to that task), and you don't get a screen forms editor for data entry and QBE
<xentrac>
I mean with the shell
<xentrac>
SQLite has a pretty decent query optimizer
<xentrac>
and of course it updates its indices properly
<xentrac>
at Satellogic we couldn't use SQLite though
<xentrac>
it didn't deal gracefully with filesystem corruption
<xentrac>
we used Postgres on the ground and a filesystem directory with JSON files on the satellites