BUG: Series.__mod__ behaves different with >1e4 rows

[river-00]

• [√] I have checked that this issue has not already been reported.

• [√] I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

# Your code here
import pandas as pd
ts = pd.Series([-2,-2,-2])
ts = pd.concat(5000 * [ts])
(ts[:10001]%4).iloc[-1]
(ts[:10000]%4).iloc[-1]

Gives
-2
2

Problem description

According to basic python module operations and previous versions of pandas, the module operation should always give a non-negative value, but when a Series larger than 10000 rows, the negative value module array operation gives a negative value.

Expected Output

2
2

Output of pd.show_versions()

1.1.1

[jbrockmendel]

series over 1e4 rows use numexpr, see core.computation.expressions._MIN_ELEMENTS. more investigation welcome

[asishm]

On 1.0.5 it outputs 2 regardless of whether numexpr is installed. On 1.1.1, with numexpr installed it returns -2 as shown (with the limit). Not sure if it would class as a regression

01623f8 is the first bad commit
commit 01623f8
Author: Martin Winkel martin.winkel.pps@gmail.com
Date: Sat Feb 1 16:13:05 2020 +0100

CLN: Refactor pandas/tests/base - part3 (#30147)