Welcome, guest | Sign In | My Account | Store | Cart

A function to create and load a table in SQLite from a Microsoft Jet table, using DAO. Also recreates indexes. Fetches and loads records in blocks with a default size of 1000 rows.

Python, 146 lines
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
"""
jetout.py - Reconstructs a given table from a Jet (MS Access) file in SQLite,
storing created table on disk; uses win32com to get to DAO 3.6 (Office 2000),
and uses sqlite3.  To ensure index names are unique to the entire database,
names of recreated indexes are prefixed with the table name.

Date/Time typed columns are reformatted to ISO's yyyy-mm-dd, and may be recalled
as datetime.date types using the built-in DATE converter in pysqlite's
converters dictionary, if the database is opened with the argument
'detect_types=sqlite3.PARSE_DECLTYPES' in the 'connect' statement.

To modify this module to import Date/Time type as datetime.datetime instead
of datetime.date:  change 'DATE' to 'TIMESTAMP' for CREATE TABLE, and modify
last line of sepdate function to include 'timepart'.

Usage:
sqlitefromjet("input.mdb", "input table name", "output.db"
[, # of rows to fetch/commit at a time])
"""
def sqlitefromjet(mdb, tbl, sqlite_db, fetchsize=1000):
        import win32com
        from win32com import client
        import sqlite3
        from itertools import imap, izip
        engine = win32com.client.Dispatch("DAO.DBEngine.36") # Office 2000
        db = engine.OpenDatabase(mdb)
        table = db.OpenRecordset(tbl)
        connection = sqlite3.connect(sqlite_db)
        cursor = connection.cursor()
        if tbl.find(' ') != -1 and tbl[:1] != '[':
                tbl = '[' + tbl + ']' # bracket table names with spaces
        # build SQL statement to create table
        createStr = 'CREATE TABLE ' + tbl + ' (' 
        fieldrange = range(table.Fields.Count)
        isType = [] # check field types (use to make ISO date from Date/Time)
        pkf = False # look for AutoNumber PK field to make INTEGER PRIMARY KEY
        havePK = False
        for tdef in db.TableDefs:
                if tdef.Name == table.Name:
                        tblDef = tdef # grab TableDef object for its indexes
                        break
        for idx in tblDef.Indexes:
                if idx.Primary:
                        for idxf in idx.Fields:
                                if pkf:
                                        pkf = False
                                        break
                                pkf = idxf.Name # grab PK field, if only one
                        break # found the PK; all done
        for field in fieldrange:
                createStr = createStr + '[' + table.Fields(field).Name + ']'
                ftype = table.Fields(field).Type # get int representing type
                if ftype == 4:
                        if pkf:
                                if table.Fields(field).Name == pkf:
                # if field was AutoNumber PK, will still autoincrement
                                        createStr += ' INTEGER PRIMARY KEY, '
                                        havePK = True
                        else:
                                createStr += ' INTEGER, '
                        isType.append(1)
                elif ftype in (2,3):
                        isType.append(1) # number
                        createStr += ' INTEGER, '
                elif ftype in (5,6,7,20):
                        if ftype == 5:
                                isType.append(3) # currency type
                        else:
                                isType.append(1) # number
                        createStr += ' NUMERIC, '
                elif ftype == 8:
                        isType.append(2) # date
                        createStr += ' DATE, ' # (or TIMESTAMP)
                else:
                        isType.append(0) # text
                        createStr += ' TEXT, '
                if table.Fields(field).Required:
                        createStr = createStr[:-2] + ' NOT NULL, '
        createStr = createStr[:-2] + ');'
        cursor.execute(createStr) # create table
        stmt = "INSERT INTO " + tbl + " VALUES(" # build INSERT for executemany
        for fieldnum in fieldrange:
                stmt += "?, "
        stmt = stmt[:-2] + ")"
        df = [x for x in fieldrange if isType[x] == 2] # date column offsets
        cf = [x for x in fieldrange if isType[x] == 3] # currency column offsets
        lastSet = False
        while not lastSet:
                # fetch 'fetchsize' records at a time (default of 1,ooo)
                fetched = map(list, table.GetRows(fetchsize)) # in mutable form
                fetchnum = len(fetched[0]) # get actual number of rows fetched
                print "\tFetched " + "%s" % fetchnum + " rows"
                # check if all 'fetchsize' rows fetched...
                if fetchnum < fetchsize:    
                        lastSet = True # ...last set (or read error) if not
                for x in df:
                        fetched[x] = imap(sepdate, fetched[x]) # yyyy-mm-dd
                for x in cf:
                        fetched[x] = imap(conv_curr, fetched[x]) # get currency
                rows = izip(*fetched) # put [fields][rows] into [rows][fields]
                cursor.executemany(stmt, rows) # INSERT one fetch
                connection.commit() # commit one transaction per fetch
        print "INSERTs complete - creating indexes..." 
        for idx in tblDef.Indexes:
                newName = table.Name + idx.Name 
                if idx.Unique:
                        if havePK and idx.Primary:
                                continue # already have unique index if int PK
                        else:
                                createStr = "CREATE UNIQUE INDEX [" + \
                                            newName + "] ON " + tbl + "("
                else:
                        createStr = "CREATE INDEX [" + newName + "] ON " + \
                                    tbl + "("
                for idxf in idx.Fields:
                        createStr += "[" + idxf.Name + "], "
                createStr = createStr[:-2] + ");"
                cursor.execute(createStr) # add an index
        print "ok - all done"
        cursor.close()
        connection.close()
        table.Close()
        db.Close()

def sepdate(dt):
        """sepdate('%m/%d/%y %H:%M:%S') -> '%Y-%m-%d'

Return ISO-format yyyy-mm-dd date from Access-formatted Date/Time."""
        dt = "%s" % dt
        if dt.find(" ") != -1:
                datepart, timepart =  dt.split(" ")
        else:
                return None
        month, day, year = datepart.split("/")
        if int(year) <= 29:
                year = '20' + year # make same assumptions about 2-digit
        elif int(year) <= 99:
                year = '19' + year # years as Access' Short Date format
        return year + "-" + month + "-" + day  # + " " + timepart # TIMESTAMP

def conv_curr(curr):
        """conv_curr(Access currency-typed field) -> float

Return a float from MS Access currency datatype, which is a fixed-point integer
scaled by 10,000"""
        return float(curr[1])/10000 # convert fixed-point int to float

To directly load a table from MS Access into SQLite, this might be handy. This solution was chosen as an alternative to exporting and loading CSV files or creating and running SQL scripts.

To get to the TableDef object and its indexing information, this code loops through the TableDefs collection to find the right TableDef. How else might a TableDef be accessed?

To modify this module to import Date/Time type as datetime.datetime instead of datetime.date: change 'DATE' to 'TIMESTAMP' for CREATE TABLE, and modify last line of separatedate function to include 'timepart'.

For information regarding pysqlite type conversion (useful for reading out date or timestamp columns created): http://initd.org/pub/software/pysqlite/doc/usage-guide.html#extensions-and-caveats http://www-eleves-isia.cma.fr/Doc/python-sqlite-1.0.1/doc/rest/manual.txt

3 comments

yannek 13 years, 4 months ago  # | flag

Very practical, thank you! But... The

while not lastSet:

in line 88ff logic is broken. Whenever a table contains a number of recordsets that is an integer multiple of "fetchsize" the while loop does not terminate before an exception occurs. What is needed is:

while not table.EOF:

And lastSet and fetchnum are then superfluous.

San Nguyen 9 years, 5 months ago  # | flag

Please check!

if ftype == 4:

   if pkf:
         if table.Fields(field).Name == pkf:
             # if field was AutoNumber PK, will still autoincrement
             createStr += ' INTEGER PRIMARY KEY, '
             havePK = True
         else: # Seems to miss ...
             createStr += ' INTEGER, ' # ... these codes!
     else:
         createStr += ' INTEGER, '
  
Hans Maurer 8 years, 11 months ago  # | flag

I've tried to convert this script with 2to3.py, but get an UnboundLocalError with tblDef which I can't fix. How do I run this in python 3.4, I am new to the language. Thx