Add an option to treat empty CSV columns as nulls, not empty strings

README.md

@@ -214,6 +214,7 @@ resolution is obtained than by using a smaller _maxzoom_ or _detail_.
    that are dropped, coalesced-as-needed, or clustered. The _operation_ may be
    `sum`, `product`, `mean`, `max`, `min`, `concat`, or `comma`
    to specify how the named _attribute_ is accumulated onto the attribute of the same name in a feature that does survive.
+ * `-pe` or `--empty-csv-columns-are-null`: Treat empty CSV columns as nulls rather than as empty strings.
 
 ### Filtering features by attributes
 
@@ -561,6 +562,7 @@ The options are:
  * `-i` or `--if-matched`: Only include features that matched the CSV.
  * `-j` *filter* or `--feature-filter`=*filter*: Check features against a per-layer filter (as defined in the [Mapbox GL Style Specification](https://www.mapbox.com/mapbox-gl-js/style-spec/#types-filter)) and only include those that match. Any features in layers that have no filter specified will be passed through. Filters for the layer `"*"` apply to all layers.
  * `-J` *filter-file* or `--feature-filter-file`=*filter-file*: Like `-j`, but read the filter from a file.
+ * `-pe` or `--empty-csv-columns-are-null`: Treat empty CSV columns as nulls rather than as empty strings.
 
 ### Setting or disabling tile size limits
 
@@ -686,6 +688,7 @@ something better.
    The formatting makes excessive use of `\u` quoting so that it follows JSON string rules but will still
    be sorted correctly by tools that just do ASCII comparisons.
  * `-c` *file.csv* or `--csv=`*file.csv*: Join attributes from the named sorted CSV file, using its first column as the join key. Geometries will be passed through even if they do not match the CSV; CSV lines that do not match a geometry will be discarded.
+ * `-pe` or `--empty-csv-columns-are-null`: Treat empty CSV columns as nulls rather than as empty strings.
 
 ### Example