From 666adc07d06958054600155930b12c3cb8dddfe3 Mon Sep 17 00:00:00 2001 From: trippy Date: Fri, 25 Nov 2022 21:58:44 +0900 Subject: [PATCH] fix code path in projection-pushdown --- user_guide/src/optimizations/lazy/projection-pushdown.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/user_guide/src/optimizations/lazy/projection-pushdown.md b/user_guide/src/optimizations/lazy/projection-pushdown.md index e2684ca07..36ebea6ab 100644 --- a/user_guide/src/optimizations/lazy/projection-pushdown.md +++ b/user_guide/src/optimizations/lazy/projection-pushdown.md @@ -27,7 +27,7 @@ Again, let's take a look the query plan. dataset.show_graph(optimized=False) ``` -![](./../outputs/projection_pushdown/graph.png) +![](../../outputs/projection_pushdown/graph.png) Now were focussed on the projection's indicated with π. The first node shows π 3/6, indicating that we select 3 out of 6 columns in the `DataFrame`. If we look the csv scans @@ -48,7 +48,7 @@ Let's see how `Polars` optimizes this query. dataset.show_graph(optimized=True) ``` -![](./../outputs/projection_pushdown/graph-optimized.png) +![](../../outputs/projection_pushdown/graph-optimized.png) The projections are pushed down the join operation all the way to the csv scans. This means that both the scanning and join operation have become cheaper due to the query