Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

import_table min_partitions broken? #5603

Closed
konradjk opened this issue Mar 14, 2019 · 4 comments

Comments

Projects
None yet
3 participants
@konradjk
Copy link
Collaborator

commented Mar 14, 2019

loading a file that has 12 partitions. when setting hl.init(min_block_size=0) and then hl.import_table(..., min_partitions=100), now getting only 3 partitions.... ☹️

@konradjk

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 21, 2019

Ok some weirdness. Running on a cluster of 2 nodes (12 cores), I get 12 partitions. With 12 nodes (92 cores), I get 4 partitions. With 22 nodes (172 cores), I get 172 partitions. (This is without a min_block_size FWIW)

@chrisvittal

This comment has been minimized.

Copy link
Collaborator

commented Mar 21, 2019

Do you have the IR with the TextTableReader json?

@konradjk

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 21, 2019

Not handy but can grab it next time I run it.

@tpoterba

This comment has been minimized.

Copy link
Collaborator

commented Mar 26, 2019

can replicate:

/Users/tpoterba/data/variants_out.txt', min_partitions=1000)
2019-03-26 09:13:47 Hail: INFO: Reading table with no type imputation
  Loading column '19:60864:GCAGCCTCAGCACT:G' as type 'str' (type not specified)


In [3]: ht.count()
[Stage 0:>                                                        (0 + 12) / 12]Out[3]: 848893

In [4]: ht.n_partitions()
Out[4]: 12

@tpoterba tpoterba assigned tpoterba and unassigned chrisvittal Mar 26, 2019

tpoterba added a commit to tpoterba/hail that referenced this issue Mar 26, 2019

danking added a commit that referenced this issue Mar 26, 2019

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.