gctools/cognos7/iqd_convert.py

from pathlib import Path
import re
import plac
from collections import defaultdict


class IqdConverter:
    commands = ["convert"]
    target_schema_by_type = {"csv": "staging", "ims": "ims"}
    output_dir = "C:\\GlobalCube\\System\\LOCOSOFT\\SQL\\schema\\LOCOSOFT\\views_imr"

    def convert(self, iqd_file, target_type="csv"):
        if not Path(iqd_file).exists():
            print(f"File {iqd_file} does not exist!")
            return
        query = self.get_query_from_iqdfile(iqd_file)
        query = self.cleanup_query(query)

        output_file = iqd_file[:-4] + ".sql"
        with open(output_file, "w", encoding="latin-1") as fwh:
            fwh.write(query)

        table = Path(iqd_file).stem.lower()
        if target_type == "csv":
            table += "_imr"
        schema = self.target_schema_by_type[target_type]

        create_view = (
            "SET QUOTED_IDENTIFIER ON \nGO\nSET ANSI_NULLS ON \nGO\n"
            + f"CREATE VIEW {schema}.{table} AS\n\n{query}\n"
            + "GO\nSET QUOTED_IDENTIFIER OFF \nGO\nSET ANSI_NULLS OFF \nGO\n\nGO"
        )

        view_file = f"{self.output_dir}\\{schema}.{table}.sql"
        with open(view_file, "w", encoding="latin-1") as fwh:
            fwh.write(create_view)

    def get_query_from_iqdfile(self, iqd_file):
        if iqd_file[-4:].lower() == ".imr":
            iqd_file = iqd_file[:-4] + ".iqd"
        iqd = open(iqd_file, "r", encoding="latin-1").read()
        res = re.findall(r"BEGIN SQL\n(.*)\n\nEND SQL", iqd, re.MULTILINE | re.DOTALL)
        query = ""
        if len(res) > 0:
            query = res[0]

        columns = re.findall(r"COLUMN,\d+,(.*)", iqd)
        col_keys = [f"c{i}" for i in range(len(columns), 0, -1)]
        col_names = [f'"{c}"' for c in reversed(columns)]

        used_cols = defaultdict(int)
        for col_key, col_name in zip(col_keys, col_names):
            used_cols[col_name] += 1
            if used_cols[col_name] > 1:
                col_name = col_name[:-1] + "_" + str(used_cols[col_name]) + '"'
            query = re.sub(col_key + r"([^\d])", col_name + r"\1", query)

        columns2 = re.findall(r'\s+(c\d+) as (".*")', query)
        used_cols = defaultdict(int)
        for col in columns2:
            used_cols[col[1]] += 1
            col_name = col[1]
            if used_cols[col[1]] > 1:
                col_name = col[1][:-1] + "_" + str(used_cols[col[1]]) + '"'
            query = re.sub(col[0] + r" as " + col[1], col_name, query)
            query = re.sub(col[0] + r"([^\d])", col_name + r"\1", query)
        return query

    def cleanup_query(self, query):
        query = re.sub(r" from (\d+) for (\d+)\)", r", \1, \2)", query)
        query = query.replace(" || ", " + ")
        query = query.replace("truncate(", "rtrim(")
        query = query.replace("char_length(", "len(")
        query = query.replace("database(", "db_name(")
        query = query.replace("ascii(", "convert(varchar(50), ")
        query = query.replace("extract(DAY FROM ", "day(")
        query = query.replace("extract(MONTH FROM ", "month(")
        query = query.replace("extract(YEAR FROM ", "year(")
        query = query.replace("od_year(", "year(")
        query = query.replace("od_month(", "month(")
        query = query.replace("lastday(", "eomonth(")
        query = query.replace("cdatetime(", "convert(datetime, ")
        query = query.replace("cast_float(", "convert(float, ")
        query = query.replace("sy_right(", "right(")
        query = query.replace("od_left(", "left(")
        query = query.replace("od_right(", "right(")
        query = query.replace("length(", "len(")
        query = query.replace("{hour}", "hour")
        query = query.replace("{minute}", "minute")
        query = query.replace("{weekday}", "weekday")
        query = query.replace("dayofweek(", "datepart(weekday, ")
        query = query.replace("cast_numberToString(cast_integer(", "((")
        query = query.replace("@CURRENT_DATE", "getdate()")
        query = query.replace("now()", "getdate()")
        query = query.replace("INTERVAL '001 10:00:00.000'", "1")
        query = query.replace("INTERVAL '001 00:00:00.000'", "1")

        query = query.replace("cdate(", "(")

        query = re.sub(r"intdiv\(([^\)]+)\,1\)", r"\1", query)
        query = re.sub(r"XCOUNT\(([^\)]+) for ", r"COUNT(\1) OVER (partition by ", query)
        query = re.sub(r"XSUM\(([^\)]+) for ", r"SUM(\1) OVER (partition by ", query)
        query = re.sub(r"RSUM\(([^\)]+) for ", r"SUM(\1) OVER (partition by ", query)
        query = re.sub(r"XMIN\(([^\)]+) for ", r"MIN(\1) OVER (partition by ", query)
        query = re.sub(r"XMAX\(([^\)]+) for ", r"MAX(\1) OVER (partition by ", query)
        query = re.sub(r"XRANK\(([^\)]+) for ([^\)]+)", r"RANK() OVER (partition by \2 order by \1", query)

        query = re.sub(r"QSS\.\"[^\"]+\\([^\\]+)\.ims\"", r'"ims"."\1"', query)
        query = re.sub(r"DATE '([\d-]+)'", r"convert(date, '\1')", query)
        query = re.sub(r"TIMESTAMP '([\d\s\-\:\.]+)'", r"convert(datetime, '\1')", query)

        query = re.sub(r"asciiz\(([^\,]*)\,\d+\)", r"convert(varchar(50), \1)", query)
        query = re.sub(r"asciiz\(([^\+]*)\)", r"convert(varchar(50), \1)", query)
        # query = re.sub(r'day\(([^\-\<\>\=]*) \- ([^\-\<\>\=]*)\)\) as', r'datediff(day, \2, \1)) as', query)
        # query = re.sub(r'day\(([^\-\<\>\=]*) \- ([^\-\<\>\=]*)\)\)\)', r'datediff(day, \2, \1)))', query)
        query = re.sub(r"day\(([^\-\,\']*) \- ", r"-1 * datediff(day, \1, ", query)
        query = re.sub(r"convert\(varchar\(50\)\, ([^,]+)\,\d+\)", r"convert(varchar(50), \1)", query)
        query = query.replace("cdate((convert(float, ", "convert(datetime, ((")

        query = re.sub(r"[^ ](order by .*)", r"\n-- \1", query)
        return query

    def run_folder(self, base_dir: str):
        files = sorted([(f.stat().st_mtime, f) for f in Path(base_dir).rglob("*.iqd")])
        for timestamp, iqd in files:
            self.convert(str(iqd))


if __name__ == "__main__":
    # IqdConverter().convert('C:\\GlobalCube_LOCOSOFT\\System\\LOCOSOFT\\IQD\\Serv_Teile\\offene_Auftraege_Ums_ben_AW.iqd')
    # IqdConverter().run_folder('C:\\Projekte\\DWH\\CARLO\\IQD')
    plac.Interpreter.call(IqdConverter)