It would be great if you could provide any details about productionizing with millions of LAS files that have 1000s of mnemonics/curves. Also, if there are any stats about load/performance testing, that would be great.
We have several Production use cases related to LAS data load, processing, and application. We used the LASIO package to process and load the data. Then we did some custom coding to streamline the process to handle any number of LAS files with any structure (not every file has the same mnemonics or the same number of mnemonics).
For example, if we process LAS log files related to MUD, we may end up seeing around 1000 distinct mnemonics, but not all these are available in every file, and saving these 1000 as 1000 columns is not a recommended approach.
We mitigated the above with our custom design and development.
I will definitely try your approach to speed up the process (if this works to productionize with no performance issues), before that, I just wanted to know your side stats about your use case, if possible.
