memory not freed reading python exported compressed feather format

[JesseTsing]

first of all, greate thanks to the authors and contributers.
I'm looking for fast data load/save solutions between python and julia for a long time, and now i've got rid of csv and database format.
i've tried a lot,but finally i choose feather/arrow. it's now the fastest format for me.

here comes the issue and reproducible codes:

here is python code creating several feather files....

import numpy as np
from pyarrow.feather import write_feather
import os

for i in range(5):
    for comp_type in ['lz4','zstd']:
        fout = 'pydata.{}.feather.{}'.format(i+1,comp_type)
        if os.path.isfile(fout):
            continue
        write_feather(pd.DataFrame(pd.np.random.rand(100000,1000)), fout, compression=comp_type)

and here is repeatly

using Arrow
using DataFrames

@showprogress 1 "read repeatly" for ii in range(1,10)
    i = mod(ii,5)+1
    @time df = copy(DataFrame(Arrow.Table("pydata.$i.feather.zstd"), copycols=true))
    finalize(df)
    sleep(1)
end

and here is the mem size ,nearly up to 16G

PID	VSZ	RSS
55623	19133564	17207488

[quinnj]

Sorry for the slow response here; I tried the code out on the latest Arrow.jl and DataFrames.jl#main and the memory seems to stay steady while reading. Can you maybe try again using the latest Arrow release and DataFrames#main to see if you can reproduce? If so, can you detail a little more what platform you're on and how you're monitoring the memory? Just want to make sure I understand exactly what you're seeing.

[JesseTsing]

YES!!! AND MANY Thanks.
when I upgrade from 1.3.0 to latest version, this error's GONE.Memory stays steady around 6G.
Maybe some dependencies are fixed. Here is my upgrading process.