Frodo | docker personal blog system | Blog library

The "_links" and "profile" are from HATEOS

It's the "Spring Data Rest" library. See https://docs.spring.io/spring-hateoas/docs/current/reference/html

maybe something like this

You can't do it. Javers' Snapshots are immutable by design.

You can zip len of word to word then create dict from len and return largest len like below:

I found the reason that because of the dic source (mysql 5.1) is too old to be compatible with the high version clickhouse. Finally I change clickhouse's version back to 19.14.6.

There are multiple ways of solving this issue, I choose what I think is the simplest approach without changing the existing pipeline too much.

I would change the second $lookup from the message to include a nested $lookup within it to fetch the user. This way we don't need to unwind the entire document and restructure the data after that.

It would look like this:

apply is inherently slower in pandas, instead of using apply we can zip the columns then inside a list comprehension test for the membership of relations_person_1 in relations_person_2 using set intersection

In general, do not use global variables. The path variable should be passed as a parameter, not kept as a global variable altered between function calls, as this leads to many side effects and is the source of countless bugs. See the following refactoring:

okey. I felt kind of stupid to make this mistake. for someone who made the same mistake, I'll make some explanation

csv - comma seperated values tsv - tab seperated values

so it's normal that cannot see tab-separated :) if you want to excel file or something create a file extension in order to that.