From: drh <> Date: Fri, 30 Jan 2026 13:11:20 +0000 (+0000) Subject: Change the new escape-character mechanism for CSV import so that there are X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=4b4df2ee1e7492fbf8413b653c5d3f8325ce4f37;p=thirdparty%2Fsqlite.git Change the new escape-character mechanism for CSV import so that there are separate options for an escape character for quoted and unquoted fields of the CSV. FossilOrigin-Name: 293ec20e29e46d822a98302f8a0ba52ad6f74a7e93f780536d393730c2edbf61 --- diff --git a/manifest b/manifest index 3cd3292f26..4ffa2758fd 100644 --- a/manifest +++ b/manifest @@ -1,5 +1,5 @@ -C Undo\sthe\sprevious\scheck-in.\s\sIn\sits\splace,\sadd\sthe\s--escape\soption\sto\nthe\s".import"\scommand. -D 2026-01-30T12:27:13.788 +C Change\sthe\snew\sescape-character\smechanism\sfor\sCSV\simport\sso\sthat\sthere\sare\nseparate\soptions\sfor\san\sescape\scharacter\sfor\squoted\sand\sunquoted\sfields\sof\nthe\sCSV. +D 2026-01-30T13:11:20.460 F .fossil-settings/binary-glob 61195414528fb3ea9693577e1980230d78a1f8b0a54c78cf1b9b24d0a409ed6a x F .fossil-settings/empty-dirs dbb81e8fc0401ac46a1491ab34a7f2c7c0452f2f06b54ebb845d024ca8283ef1 F .fossil-settings/ignore-glob 35175cdfcf539b2318cb04a9901442804be81cd677d8b889fcc9149c21f239ea @@ -739,7 +739,7 @@ F src/random.c 606b00941a1d7dd09c381d3279a058d771f406c5213c9932bbd93d5587be4b9c F src/resolve.c 47aa7fdc9ec4c19b103ac5e79d7887d30119b5675309facf5eed1118391c868b F src/rowset.c 8432130e6c344b3401a8874c3cb49fefe6873fec593294de077afea2dce5ec97 F src/select.c 4bee1bb231771e7c6e5aef243b1f74c3d330df5a005909d5e2c338fb1510fe55 -F src/shell.c.in 9f45c5d4b25c4621dc47cee7a3388ff1d8e8d1931240d4bfb60332bd3710e679 +F src/shell.c.in 461c163897e1d001e957c69cad624ddac367e24e6da0b2de54592ea4e18fe98a F src/sqlite.h.in 8bcbaecfe2cbecf8c5c1381354fcdd7d307443e88b4953fccb222456c1267b61 F src/sqlite3.rc 015537e6ac1eec6c7050e17b616c2ffe6f70fca241835a84a4f0d5937383c479 F src/sqlite3ext.h 1b7a0ee438bb5c2896d0609c537e917d8057b3340f6ad004d2de44f03e3d3cca @@ -1621,7 +1621,7 @@ F test/shell1.test 56f200ba4b36cbe8229d84d848e6a8d69eb91a75b2deaf28c454027433c6e F test/shell2.test dc541d2681503e55466a24d35a4cbf8ca5b90b8fcdef37fc4db07373a67d31d3 F test/shell3.test 603b448e917537cf77be0f265c05c6f63bc677c63a533c8e96aae923b56f4a0e F test/shell4.test e25580a792b7b54560c3a76b6968bd8189261f38979fe28e6bc6312c5db280db -F test/shell5.test d2aee9c19b6aa0e4bde599e194cd01df60c97f5b721d6493cf8c1da0b45cede5 +F test/shell5.test c5e8fb055ca74511eb0e757b568e48853d99306aea4a7e6a47a61dcf492a7d3c F test/shell6.test e3b883b61d4916b6906678a35f9d19054861123ad91b856461e0a456273bdbb8 F test/shell7.test 43fd8e511c533bab5232e95c7b4be93b243451709e89582600d4b6e67693d5c3 F test/shell8.test 641cf21a99c59404c24e3062923734951c4099a6b6b6520de00cf7a1249ee871 @@ -2193,8 +2193,8 @@ F tool/warnings-clang.sh bbf6a1e685e534c92ec2bfba5b1745f34fb6f0bc2a362850723a9ee F tool/warnings.sh d924598cf2f55a4ecbc2aeb055c10bd5f48114793e7ba25f9585435da29e7e98 F tool/win/sqlite.vsix deb315d026cc8400325c5863eef847784a219a2f F tool/winmain.c 00c8fb88e365c9017db14c73d3c78af62194d9644feaf60e220ab0f411f3604c -P 7ba0594d3c24531bc014dd029093b0503118a29cfdd13af2a2ebdd456855ce94 -R 6f4df6c212e3dffec0985d0418b3eb03 +P f11a05d52a1eb933f394d48c0d4ee0353f07ad725b7378c40afd01632e67f72d +R a8ba995b68957165836a359cdc9bda8c U drh -Z 9898050fb5f26c38e6d09b053652c74e +Z ebb3f29c37b6be76dd377e9cc977e6a7 # Remove this line to create a well-formed Fossil manifest. diff --git a/manifest.uuid b/manifest.uuid index 8f02d1ad8d..9910be4c8d 100644 --- a/manifest.uuid +++ b/manifest.uuid @@ -1 +1 @@ -f11a05d52a1eb933f394d48c0d4ee0353f07ad725b7378c40afd01632e67f72d +293ec20e29e46d822a98302f8a0ba52ad6f74a7e93f780536d393730c2edbf61 diff --git a/src/shell.c.in b/src/shell.c.in index 846bf58531..d9b32e4859 100644 --- a/src/shell.c.in +++ b/src/shell.c.in @@ -4920,7 +4920,8 @@ struct ImportCtx { int cTerm; /* Character that terminated the most recent field */ int cColSep; /* The column separator character. (Usually ",") */ int cRowSep; /* The row separator character. (Usually "\n") */ - int cEscape; /* Escape character. 0 for none */ + int cQEscape; /* Escape character with "...". 0 for none */ + int cUQEscape; /* Escape character not with "...". 0 for none */ }; /* Clean up resourced used by an ImportCtx */ @@ -4979,7 +4980,6 @@ static char *SQLITE_CDECL csv_read_one_field(ImportCtx *p){ int c; int cSep = (u8)p->cColSep; int rSep = (u8)p->cRowSep; - int cEsc = (u8)p->cEscape; p->n = 0; c = import_getc(p); if( c==EOF || seenInterrupt ){ @@ -4990,6 +4990,7 @@ static char *SQLITE_CDECL csv_read_one_field(ImportCtx *p){ int pc, ppc; int startLine = p->nLine; int cQuote = c; + int cEsc = (u8)p->cQEscape; pc = ppc = 0; while( 1 ){ c = import_getc(p); @@ -4997,8 +4998,7 @@ static char *SQLITE_CDECL csv_read_one_field(ImportCtx *p){ if( c==cEsc && cEsc!=0 ){ c = import_getc(p); import_append_char(p, c); - ppc = pc; - pc = c; + ppc = pc = 0; continue; } if( c==cQuote ){ @@ -5033,6 +5033,7 @@ static char *SQLITE_CDECL csv_read_one_field(ImportCtx *p){ }else{ /* If this is the first field being parsed and it begins with the ** UTF-8 BOM (0xEF BB BF) then skip the BOM */ + int cEsc = p->cUQEscape; if( (c&0xff)==0xef && p->bNotFirst==0 ){ import_append_char(p, c); c = import_getc(p); @@ -7411,7 +7412,8 @@ static int pickStr(const char *zArg, char **pzErr, ...){ ** delimiters are specified using --colsep and/or --rowsep ** --colsep CHAR Use CHAR as the column separator. ** --csv Input is standard RFC-4180 CSV. -** --escape CHAR Use CHAR as an escape character in CSV imports. +** --esc CHAR Use CHAR as an escape character in unquoted CSV inputs. +** --qesc CHAR Use CHAR as an escape character in quoted CSV inputs. ** --rowsep CHAR Use CHAR as the row separator. ** --schema S When creating TABLE, put it in schema S ** --skip N Ignore the first N rows of input @@ -7470,8 +7472,10 @@ static int dotCmdImport(ShellState *p){ if( sCtx.cColSep==0 ) sCtx.cColSep = ','; if( sCtx.cRowSep==0 ) sCtx.cRowSep = '\n'; xRead = csv_read_one_field; - }else if( cli_strcmp(z,"-escape")==0 ){ - sCtx.cEscape = azArg[++i][0]; + }else if( cli_strcmp(z,"-esc")==0 ){ + sCtx.cUQEscape = azArg[++i][0]; + }else if( cli_strcmp(z,"-qesc")==0 ){ + sCtx.cQEscape = azArg[++i][0]; }else if( cli_strcmp(z,"-colsep")==0 ){ if( i==nArg-1 ){ dotCmdError(p, i, "missing argument", 0); diff --git a/test/shell5.test b/test/shell5.test index 3ba8df463d..b21b24b924 100644 --- a/test/shell5.test +++ b/test/shell5.test @@ -344,16 +344,16 @@ do_test shell5-1.10.1 { set out [open shell5.csv w] fconfigure $out -translation lf puts $out {column1,column2,column3,column4} - puts $out "x1,x2,\"x3 \\\"data\\\" 3\",x4" + puts $out "x1,x2%\"x3,\"x3\\\"data\\\"3\",x4" close $out forcedelete test.db catchcmd test.db { CREATE TABLE t1(a,b,c,d); -.import --csv --escape \\ shell5.csv t1 +.import --csv --qesc \\ --esc % shell5.csv t1 } sqlite3 db test.db - db eval {SELECT c FROM t1 ORDER BY rowid} -} {column3 {x3 "data" 3}} + db eval {SELECT b, c FROM t1 ORDER BY rowid} +} {column2 column3 x2\"x3 x3\"data\"3} # Blank last column with \r\n line endings. do_test shell5-1.11 {