Globaldb migration: new tables and move assets #3400
Conversation
| @@ -622,7 +622,7 @@ def __post_init__( | |||
| object.__setattr__(self, 'decimals', data.decimals) | |||
| object.__setattr__(self, 'protocol', data.protocol) | |||
|
|
|||
| underlying_tokens = GlobalDBHandler().fetch_underlying_tokens(data.ethereum_address) | |||
| underlying_tokens = GlobalDBHandler().fetch_underlying_tokens(data.identifier) | |||
There was a problem hiding this comment.
Also above here: https://github.com/rotki/rotki/pull/3400/files#diff-d838b4670ae94dd2743ece1ac79b73a99865b761474821ed39d7d3498d830338R616
Shouldn't it be talking about an EVM token?
There was a problem hiding this comment.
Need to rename some things. Will do in the next PR
| return {k.value: addr for k in CHAINID} | ||
|
|
||
|
|
||
| def query_coingecko_for_addresses(identifier: str) -> Dict[str, str]: # noqa: E501 pylint: disable=unused-argument |
There was a problem hiding this comment.
As discussed. This CAN NOT be part of the actual DB migration. This can only happen after the migration is complete, the user has logged in and a long-running task can executed.
There was a problem hiding this comment.
Yes I agree. Won't use it right here but needed a function to query ids. Will move this function to a different place since I don't use it
|
|
||
| for entry in query: | ||
| try: | ||
| ids = query_coingecko_for_addresses(entry[6]) |
There was a problem hiding this comment.
Querying any external service, especially in a for loop should not be part of a DB upgrade, ever. A DB upgrade is one of the most sensitive things we can do and if it goes wrong the user's installation may be fucked. So we need to keep it as simple as possible.
There was a problem hiding this comment.
yeah. This logic has to move but it will be in another PR. They ids and information would come from the packaged globaldb but at this point I needed something to simulate this and test that the migration works. Also will be useful to move it into the script we will use to create the new globaldb
yabirgb
force-pushed
the
gdb_v2_tables
branch
from
48f27f5
to
1bd404a
Compare
yabirgb
force-pushed
the
gdb_v2_tables
branch
from
1bd404a
to
f966b58
Compare
| @@ -148,6 +160,49 @@ | |||
| ); | |||
| """ | |||
|
|
|||
| DB_V3_CREATE_COMMON_ASSETS_DETAILS = """ | |||
There was a problem hiding this comment.
And this won't stay right? We willl only have 1 version of creation command per table.
There was a problem hiding this comment.
I was thinking in someone comming from an old version but I believe it doesn't make sense? We can go ahead and remove the versions in the names
There was a problem hiding this comment.
it doesn't make sense. We never keep old version creation schemas around. That's what DB upgrades are for.
Danger
This branch will modify your globaldb and break it. Make a copy of it before executing ANY code of it.
What has been done