From d58a0f5e2899af49e922f1a00f8313b2429f9489 Mon Sep 17 00:00:00 2001
From: Xixiang-Liu <xixiangliuoscar@gmail.com>
Date: Mon, 1 Apr 2024 20:07:51 -0500
Subject: [PATCH] get rid of warnings

---
 example.parquet | Bin 2169 -> 2169 bytes
 parquet.py      |  11 +++++++----
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/example.parquet b/example.parquet
index e56baf7c7fb63875721ad14164a93c1c9b540d2e..647727c41f2eeba762fb450ddb8297a5a8c8c215 100644
GIT binary patch
delta 332
zcmWN@TQWic007X;ko$9IZiw;<{rISeP~`nqDR1(&5nHGw_^^UDV=MhRA7|$Dqkhy`
zObmuWY=}6o1l>xe5lNX5Mj0c?I1{8u(_o5eW|-xKIp$d)+x^Xc=Oj;oC6+0&!YU=!
zSZ9L-6*k#opKW&7WsfQc9CE}l8BRIloC|7U(2mt*l1rMjxZ;`{Zn@*02OfE%%`-0y
S^Ts<Le1<{wtM@bT7yk#vtRP_k

delta 332
zcmWN@TQWic007X;ko$9IZiw;<{rISeP~`nqDR1(&5nHGw_^^UDV=MhRA7|$Dqkhy`
zOr*mgHbk6ZM!J<uV^R_Xj50=&874?ENW>J=j5AAy73NuGtNWY%-jqBAmRP3991E;b
zVx0}LRLD`L$^$#>vd0w%9CE}lC!BJ|Ef-vdK|5BL1~r<r*yoxX&bi~BZ60}|%`-2&
S^2R$~e1<{wqxUoL7yk!B2q0kq

diff --git a/parquet.py b/parquet.py
index 04d9a0a..d66e623 100644
--- a/parquet.py
+++ b/parquet.py
@@ -49,14 +49,17 @@ def degrade_sortedness_to_target(df: pd.DataFrame) -> None:
     # swap random value pairs in the block till reach target_sortedness
     def degrade_block(block: pd.DataFrame) -> None:
         N = len(block)
+        index_values = block.index.tolist()
         # prevent infinite loop in case target can't be reach
         for _ in range(10000):
             sortedness_block = get_sortedness_block(block)
             if sortedness_block <= target_sortedness:
                 return
-            idx1 = random.randint(0, N - 1)
-            idx2 = random.randint(0, N - 1)
-            block.iloc[idx1], block.iloc[idx2] = block.iloc[idx2], block.iloc[idx1].copy()
+            idx1 = random.choice(index_values)
+            idx2 = random.choice(index_values)
+            tmp = block['col'][idx1].copy()
+            block.at[idx1, 'col'] = block.at[idx2, 'col']
+            block.at[idx2, 'col'] = tmp
 
     num_rows = len(df)
     num_full_blocks = num_rows // size_block
@@ -78,7 +81,7 @@ def degrade_block(block: pd.DataFrame) -> None:
 target_sortedness = 0.8
 size_block = 512
 
-df = pd.DataFrame({"col": range(1000)})
+df = pd.DataFrame({"col": range(100)})
 degrade_sortedness_to_target(df)
 table = pa.Table.from_pandas(df)
 pq.write_table(table, "example.parquet")