From ec193ca63999451abea20592a60804f85781cc57 Mon Sep 17 00:00:00 2001
From: Daniil Medvedev <medvdanil@gmail.com>
Date: Fri, 17 Jul 2015 11:48:33 +0300
Subject: [PATCH] csv output(with tests)
---
src/lib/csv/csv.c | 23 +++++++++++++-
src/lib/csv/csv.h | 15 +++++-----
src/lua/csv.lua | 71 +++++++++++++++++++++++++++++++-------------
test/unit/csv.c | 18 +++++++++++
test/unit/csv.result | 3 ++
5 files changed, 102 insertions(+), 28 deletions(-)
diff --git a/src/lib/csv/csv.c b/src/lib/csv/csv.c
index d4856b97f0..25a8050eb4 100644
--- a/src/lib/csv/csv.c
+++ b/src/lib/csv/csv.c
@@ -242,7 +242,28 @@ csv_next(struct csv_iterator *it) {
}
void
-csv_feed(struct csv_iterator *it, const char *str) {
+csv_feed(struct csv_iterator *it, const char *str)
+{
it->buf_begin = str;
it->buf_end = str + strlen(str);
}
+
+int
+csv_escape_field(struct csv *csv, const char *field, char *dst)
+{
+ char *p = dst;
+ int inquotes = 0;
+ if(strchr(field, csv->csv_delim) || strchr(field, '\n') || strchr(field, '\r')) {
+ inquotes = 1;
+ *p++ = csv->csv_quote;
+ }
+ while(*field) {
+ if(*field == csv->csv_quote)
+ *p++ = csv->csv_quote;
+ *p++ = *field++;
+ }
+ if(inquotes)
+ *p++ = csv->csv_quote;
+ *p = 0;
+ return p - dst;
+}
diff --git a/src/lib/csv/csv.h b/src/lib/csv/csv.h
index bfb699e645..fb2e9376a6 100644
--- a/src/lib/csv/csv.h
+++ b/src/lib/csv/csv.h
@@ -81,13 +81,6 @@ csv_parse_chunk(struct csv *csv, const char *s, const char *end);
void
csv_finish_parsing(struct csv *csv);
-/**
- * Format variadic arguments and print them into
- * a stream, adding CSV markup.
- */
-int
-csv_snprintf(struct csv *csv, FILE *f, const char *format, ...);
-
/**
* if quote not closed returns 0
*/
@@ -122,6 +115,14 @@ csv_next(struct csv_iterator *);
void
csv_feed(struct csv_iterator *, const char *);
+/**
+ * @brief csv_escape_field adds pair quote and
+ * if there is comma in field, adds surrounding quotes
+ */
+
+int
+csv_escape_field(struct csv *csv, const char *field, char *dst);
+
#define CSV_ITERATOR_GET_FIELD(it) it->field
#define CSV_ITERATOR_GET_FLEN(it) it->field_len
diff --git a/src/lua/csv.lua b/src/lua/csv.lua
index 13f437ce84..92ac72a334 100644
--- a/src/lua/csv.lua
+++ b/src/lua/csv.lua
@@ -41,20 +41,21 @@ ffi.cdef[[
void csv_iter_create(struct csv_iterator *it, struct csv *csv);
int csv_next(struct csv_iterator *);
void csv_feed(struct csv_iterator *, const char *);
+ int csv_escape_field(struct csv *, const char *, char *);
enum {
- CSV_IT_OK,
- CSV_IT_EOL,
- CSV_IT_NEEDMORE,
- CSV_IT_EOF,
- CSV_IT_ERROR
+ CSV_IT_OK,
+ CSV_IT_EOL,
+ CSV_IT_NEEDMORE,
+ CSV_IT_EOF,
+ CSV_IT_ERROR
};
]]
csv = {
-loadcsv = function(readable, csv_chunk_size)
+load = function(readable, csv_chunk_size)
csv_chunk_size = csv_chunk_size or 4096
if type(readable.read) ~= "function" then
- error("Usage: loadcsv(object with read method)")
+ error("Usage: load(object with read method)")
end
local log = require('log')
@@ -66,23 +67,53 @@ loadcsv = function(readable, csv_chunk_size)
local tup = {}
local result = {}
while st ~= ffi.C.CSV_IT_EOF do
- if st == ffi.C.CSV_IT_NEEDMORE then
- ffi.C.csv_feed(it, readable:read(csv_chunk_size))
- elseif st == ffi.C.CSV_IT_EOL then
- table.insert(result, tup)
- tup = {}
- elseif st == ffi.C.CSV_IT_OK then
- table.insert(tup, ffi.string(it[0].field, it[0].field_len))
- elseif st == ffi.C.CSV_IT_ERROR then
- log.warn("CSV file has errors")
- elseif st == ffi.C.CSV_IT_EOF then
- break
- end
- st = ffi.C.csv_next(it)
+ if st == ffi.C.CSV_IT_NEEDMORE then
+ ffi.C.csv_feed(it, readable:read(csv_chunk_size))
+ elseif st == ffi.C.CSV_IT_EOL then
+ table.insert(result, tup)
+ tup = {}
+ elseif st == ffi.C.CSV_IT_OK then
+ table.insert(tup, ffi.string(it[0].field, it[0].field_len))
+ elseif st == ffi.C.CSV_IT_ERROR then
+ log.warn("CSV file has errors")
+ elseif st == ffi.C.CSV_IT_EOF then
+ break
+ end
+ st = ffi.C.csv_next(it)
end
return result
end
+,
+dump = function(writable, t)
+ if type(writable.write) ~= "function" or type(t) ~= "table" then
+ error("Usage: dump(writable, table)")
+ end
+ local csv = ffi.new('csv_t[1]')
+ ffi.C.csv_create(csv)
+ local bufsz = 256
+ --local buf = ffi.new('char[?]', bufsz)
+ local buf = csv[0].csv_realloc(ffi.cast(ffi.typeof('void *'), 0), bufsz)
+ for k, line in pairs(t) do
+ first = true
+ for k2, field in pairs(line) do
+ strf = tostring(field)
+ if (strf:len() + 1) * 2 > bufsz then
+ bufsz = (strf:len() + 1) * 2
+ buf = csv[0].csv_realloc(buf, bufsz)
+ end
+ local len = ffi.C.csv_escape_field(csv, strf, buf)
+ if first then
+ first = false
+ else
+ writable:write(',')
+ end
+ writable:write(ffi.string(buf, len))
+ end
+ writable:write('\n')
+ end
+ csv[0].csv_realloc(buf, 0)
+end
}
return csv
diff --git a/test/unit/csv.c b/test/unit/csv.c
index 2386568f46..8a36c7b6b6 100644
--- a/test/unit/csv.c
+++ b/test/unit/csv.c
@@ -247,6 +247,22 @@ void iter_test2() {
footer();
}
+void csv_out() {
+ header();
+
+ const char fields[4][36] = { "abc", "with,comma", "\"in quotes\"", "1 \" quote"};
+ char buf[18];
+ int i;
+ struct csv csv;
+ csv_create(&csv);
+ for(i = 0; i < 4; i++) {
+ int len = csv_escape_field(&csv, fields[i], buf);
+ printf("%s<len=%d>%c", buf, len, i == 3 ? '\n' : ',');
+ }
+
+ footer();
+}
+
int main() {
test1();
test2();
@@ -314,5 +330,7 @@ int main() {
iter_test1();
iter_test2();
+ //output test
+ csv_out();
return 0;
}
diff --git a/test/unit/csv.result b/test/unit/csv.result
index 9e43199617..8d56fa9744 100644
--- a/test/unit/csv.result
+++ b/test/unit/csv.result
@@ -183,4 +183,7 @@ valid: yes
|1|
|23|
*** iter_test2: done ***
+ *** csv_out ***
+abc<len=3>,"with,comma"<len=12>,""in quotes""<len=13>,1 "" quote<len=10>
+ *** csv_out: done ***
\ No newline at end of file
--
GitLab