Starts sparkxgb section

_posts/2024-04-22-sparklyr-updates/sparklyr-updates.Rmd

@@ -28,6 +28,9 @@ knitr::opts_chunk$set(
 
 ## Highlights
 
+`sparklyr` and friends have been getting some important updates in the past few
+months, here are some highlights:
+
 * Databricks Connect v2 now supports running native R code in Spark, via
 `pysparklyr`. 
 
@@ -67,27 +70,25 @@ for the next time you run the call.
 A full article about this new capability is available here: 
 [Run R inside Databricks Connect](https://spark.posit.co/deployment/databricks-connect-udfs.html)
 
-## sparklyr 1.8.5
-
-### Fixes
-
-- Fixes quoting issue with `dbplyr` 2.5.0 (#3429)
-
-- Fixes Windows OS identification (#3426)
+## sparkxgb 
 
-### Package improvements
+The `sparkxgb` is an extension of `sparklyr`. It enables integration with 
+[XGBoost](https://xgboost.readthedocs.io/en/stable/). The current CRAN release
+does not support the more recent versions of XGBoost. This limitation has recently 
+prompted a full refresh of `sparkxgb`. Here is a summary of the improvements,
+currently in the [development version of the package](https://github.com/rstudio/sparkxgb): 
 
-- Removes dependency on `tibble`, all calls are now redirected to `dplyr` (#3399)
+- The `xgboost_classifier()` and `xgboost_regressor()` functions no longer
+pass values from two existing arguments:
 
-- Removes dependency on `rapddirs` (#3401): 
-  - Backwards compatibility with `sparklyr` 0.5 is no longer needed
-  - Replicates selection of cache directory 
+  - `sketch_eps`
+  - `timeout_request_workers`
 
-- Converts `spark_apply()` to a method (#3418)
+These are parameters that XGBoost has deprecated. This is the main reason why
+R users are seeing errors when using the version currently on CRAN. 
 
-## sparkxgb 
 
-- Avoids sending two deprecated parameters to XGBoost. The default arguments in
+Avoids sending two deprecated parameters to XGBoost. The default arguments in
 the R function are NULL, and it will return an error message if the call intends
 to use them:
 
@@ -118,3 +119,20 @@ Spark session
   package moving forward.
 
 
+## sparklyr 1.8.5
+
+### Fixes
+
+- Fixes quoting issue with `dbplyr` 2.5.0 (#3429)
+
+- Fixes Windows OS identification (#3426)
+
+### Package improvements
+
+- Removes dependency on `tibble`, all calls are now redirected to `dplyr` (#3399)
+
+- Removes dependency on `rapddirs` (#3401): 
+  - Backwards compatibility with `sparklyr` 0.5 is no longer needed
+  - Replicates selection of cache directory 
+
+- Converts `spark_apply()` to a method (#3418)