WIP: Support converting CSV with JSON metadata to sas7bdat

.gitignore

@@ -45,9 +45,12 @@ gen.json
 extract_metadata
 test_sav_date
 test_double_decimals
+dump_sas7bdat
 generate_corpus
 *.csv
 *.json
+!resources/csv_to_sas7bdat.csv
+!resources/csv_to_sas7bdat.json
 *.py
 *.tar.gz
 *.zip

Makefile.am

@@ -95,7 +95,11 @@ EXTRA_DIST = \
 	resources/compression_type_0x0d.sas7bdat \
 	resources/format_with_default_32bit.sas7bcat \
 	resources/format_with_default_64bit.sas7bcat \
-	resources/sav_missing_continuations.sav
+	resources/sav_missing_continuations.sav \
+	resources/csv_to_sas7bdat.csv \
+	resources/csv_to_sas7bdat.json \
+	resources/csv_to_sas7bdat.expected \
+	src/test/test_csv_to_sas7bdat.sh
 
 include_HEADERS = src/readstat.h
 
@@ -137,6 +141,7 @@ noinst_HEADERS = \
        src/bin/read_csv/mod_csv.h \
        src/bin/read_csv/mod_dta.h \
        src/bin/read_csv/mod_sav.h \
+       src/bin/read_csv/mod_sas7bdat.h \
        src/bin/read_csv/read_csv.h \
        src/bin/read_csv/read_module.h \
        src/bin/read_csv/value.h \
@@ -177,6 +182,7 @@ readstat_SOURCES = \
 	src/bin/read_csv/mod_csv.c \
 	src/bin/read_csv/mod_dta.c \
 	src/bin/read_csv/mod_sav.c \
+	src/bin/read_csv/mod_sas7bdat.c \
 	src/bin/read_csv/value.c \
 	src/bin/write/double_decimals.c \
 	src/bin/write/mod_csv.c \
@@ -260,6 +266,16 @@ test_double_decimals_CFLAGS = -g -Wall @EXTRA_WARNINGS@ -Werror -pedantic-errors
 
 TESTS = test_readstat test_dta_days test_sav_date test_double_decimals
 
+if HAVE_CSVREADER
+check_PROGRAMS += dump_sas7bdat
+
+dump_sas7bdat_SOURCES = src/test/dump_sas7bdat.c
+dump_sas7bdat_LDADD = libreadstat.la
+dump_sas7bdat_CFLAGS = -Wall @EXTRA_WARNINGS@ -Werror -pedantic-errors -std=c99
+
+TESTS += src/test/test_csv_to_sas7bdat.sh
+endif
+
 EXTRA_PROGRAMS = \
     generate_corpus
 

README.md

@@ -138,7 +138,9 @@ CSV input is supported together with a metadata file describing the data:
 
     readstat <input file.csv> <input metadata.json> <output file>
 
-The `<output file>` should end with `.dta`, `.sav`, or `.csv`.
+The `<output file>` should end with `.dta`, `.sav`, `.sas7bdat`, or `.csv`.
+A `.sas7bdat` file cannot hold value labels, which SAS keeps in a separate
+catalog, so any `categories` in the metadata are not written to it.
 
 The `<input file.csv>` is a regular CSV file.
 

VS17/ReadStat_App/ReadStat_App.vcxproj

@@ -25,6 +25,7 @@
     <ClCompile Include="..\..\src\bin\read_csv\mod_csv.c" />
     <ClCompile Include="..\..\src\bin\read_csv\mod_dta.c" />
     <ClCompile Include="..\..\src\bin\read_csv\mod_sav.c" />
+    <ClCompile Include="..\..\src\bin\read_csv\mod_sas7bdat.c" />
     <ClCompile Include="..\..\src\bin\read_csv\value.c" />
     <ClCompile Include="..\..\src\bin\util\readstat_dta_days.c" />
     <ClCompile Include="..\..\src\bin\util\readstat_sav_date.c" />
@@ -43,6 +44,7 @@
     <ClInclude Include="..\..\src\bin\read_csv\mod_csv.h" />
     <ClInclude Include="..\..\src\bin\read_csv\mod_dta.h" />
     <ClInclude Include="..\..\src\bin\read_csv\mod_sav.h" />
+    <ClInclude Include="..\..\src\bin\read_csv\mod_sas7bdat.h" />
     <ClInclude Include="..\..\src\bin\read_csv\read_csv.h" />
     <ClInclude Include="..\..\src\bin\read_csv\read_module.h" />
     <ClInclude Include="..\..\src\bin\read_csv\value.h" />

VS17/ReadStat_App/ReadStat_App.vcxproj.filters

@@ -45,6 +45,9 @@
     <ClCompile Include="..\..\src\bin\read_csv\mod_sav.c">
       <Filter>Source Files</Filter>
     </ClCompile>
+    <ClCompile Include="..\..\src\bin\read_csv\mod_sas7bdat.c">
+      <Filter>Source Files</Filter>
+    </ClCompile>
     <ClCompile Include="..\..\src\bin\read_csv\value.c">
       <Filter>Source Files</Filter>
     </ClCompile>
@@ -95,6 +98,9 @@
     <ClInclude Include="..\..\src\bin\read_csv\mod_sav.h">
       <Filter>Header Files</Filter>
     </ClInclude>
+    <ClInclude Include="..\..\src\bin\read_csv\mod_sas7bdat.h">
+      <Filter>Header Files</Filter>
+    </ClInclude>
     <ClInclude Include="..\..\src\bin\read_csv\read_csv.h">
       <Filter>Header Files</Filter>
     </ClInclude>

resources/csv_to_sas7bdat.csv

@@ -0,0 +1,5 @@
+id,amount,pct,score,born,stamp,clock,name
+1,12.5,25,3,2020-02-29,2020-02-29 13:45:30.250,08:15:00,Zoë
+2,,50,-9,1959-12-31,1960-01-01T00:00:01,23:59:59.5,
+3,-0.75,,999,,2016-12-31 23:59:59,,"a, b"
+4,1e3,100,7,1960-01-01,,00:00:00,plain

resources/csv_to_sas7bdat.expected

@@ -0,0 +1,40 @@
+variable id format=BEST12. label=Identifier
+variable amount format=F12.2 label=
+variable pct format=F11.1 label=
+variable score format=BEST12. label=
+variable born format=DATE9. label=
+variable stamp format=DATETIME19. label=
+variable clock format=TIME8. label=
+variable name format= label=Name
+0 id 1
+0 amount 12.5
+0 pct 25
+0 score 3
+0 born 21974
+0 stamp 1898603130.25
+0 clock 29700
+0 name "Zoë"
+1 id 2
+1 amount .
+1 pct 50
+1 score .A
+1 born .A
+1 stamp 1
+1 clock 86399.5
+1 name ""
+2 id 3
+2 amount -0.75
+2 pct .
+2 score .B
+2 born .
+2 stamp 1798847999
+2 clock .
+2 name "a, b"
+3 id 4
+3 amount 1000
+3 pct 100
+3 score 7
+3 born 0
+3 stamp .
+3 clock 0
+3 name "plain"

resources/csv_to_sas7bdat.json

@@ -0,0 +1,11 @@
+{"type": "SAS", "variables": [
+ {"type": "NUMERIC", "name": "id", "label": "Identifier"},
+ {"type": "NUMERIC", "name": "amount", "format": "CURRENCY", "decimals": 2},
+ {"type": "NUMERIC", "name": "pct", "format": "PERCENT", "decimals": 1},
+ {"type": "NUMERIC", "name": "score", "missing": {"type": "DISCRETE", "values": [-9, 999]},
+   "categories": [{"code": 3, "label": "Three"}]},
+ {"type": "NUMERIC", "name": "born", "format": "DATE", "missing": {"type": "RANGE", "low": "1959-01-01", "high": "1959-12-31"}},
+ {"type": "NUMERIC", "name": "stamp", "format": "DATE_TIME"},
+ {"type": "NUMERIC", "name": "clock", "format": "TIME"},
+ {"type": "STRING", "name": "name", "label": "Name"}
+]}