for compress the offsets before delta

spiraldb · Sep 30, 2024 · b50d382 · b50d382
1 parent f7636d9
commit b50d382
Showing 1 changed file with 11 additions and 3 deletions.
diff --git a/vortex-sampling-compressor/src/compressors/fsst.rs b/vortex-sampling-compressor/src/compressors/fsst.rs
@@ -12,6 +12,7 @@ use vortex_error::{vortex_bail, VortexResult};
 use vortex_fsst::{fsst_compress, fsst_train_compressor, FSSTArray, FSSTEncoding, FSST};
 
 use super::delta::DeltaCompressor;
+use super::r#for::FoRCompressor;
 use super::{CompressedArray, CompressionTree, EncoderMetadata, EncodingCompressor};
 use crate::SamplingCompressor;
 
@@ -96,15 +97,21 @@ impl EncodingCompressor for FSSTCompressor {
         let codes_varbin = VarBinArray::try_from(fsst_array.codes())?;
         let codes_varbin_dtype = codes_varbin.dtype().clone();
 
-        let codes_offsets_delta = DeltaCompressor.compress(
+        let codes_offsets_for = FoRCompressor.compress(
             &codes_varbin.offsets(),
             like.as_ref().and_then(|l| l.child(1).cloned()),
-            ctx.auxiliary("offsets"),
+            ctx.auxiliary("offsets_for"),
+        )?;
+
+        let codes_offsets_delta = DeltaCompressor.compress(
+            &codes_offsets_for.array,
+            like.as_ref().and_then(|l| l.child(2).cloned()),
+            ctx.auxiliary("offsets_for_delta"),
         )?;
 
         let codes_offsets_compressed = codes_compressor.auxiliary("delta_offsets").compress(
             &codes_offsets_delta.array,
-            like.as_ref().and_then(|l| l.child(2)),
+            like.as_ref().and_then(|l| l.child(3)),
         )?;
 
         let codes = VarBinArray::try_new(
@@ -128,6 +135,7 @@ impl EncodingCompressor for FSSTCompressor {
                 self,
                 vec![
                     uncompressed_lengths.path,
+                    codes_offsets_for.path,
                     codes_offsets_delta.path,
                     codes_offsets_compressed.path,
                 ],