bandas-df 0.3.0__tar.gz

bandas_df-0.3.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 salim-studio
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

bandas_df-0.3.0/PKG-INFO

@@ -0,0 +1,246 @@
+Metadata-Version: 2.4
+Name: bandas-df
+Version: 0.3.0
+Summary: bandas — pandas-compatible, faster DataFrame + universal data platform (SQL/DB, time-series, preprocessing, ML/DL, viz, lazy)
+License: MIT
+Keywords: dataframe,pandas,fast,numpy,sql,database,machine-learning,deep-learning,data-science,analytics,time-series,preprocessing
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.26
+Requires-Dist: pandas>=2.0
+Requires-Dist: numba>=0.57
+Requires-Dist: pyarrow>=12
+Requires-Dist: sqlalchemy>=2.0
+Requires-Dist: duckdb>=0.9
+Requires-Dist: scikit-learn>=1.3
+Requires-Dist: scipy>=1.10
+Requires-Dist: matplotlib>=3.7
+Requires-Dist: seaborn>=0.12
+Requires-Dist: openpyxl>=3.1
+Requires-Dist: xlrd>=2.0
+Provides-Extra: speed
+Requires-Dist: numba>=0.57; extra == "speed"
+Requires-Dist: pyarrow>=12; extra == "speed"
+Provides-Extra: db
+Requires-Dist: sqlalchemy>=2.0; extra == "db"
+Requires-Dist: duckdb>=0.9; extra == "db"
+Provides-Extra: ml
+Requires-Dist: scikit-learn>=1.3; extra == "ml"
+Requires-Dist: scipy>=1.10; extra == "ml"
+Provides-Extra: dl
+Requires-Dist: torch>=2.0; extra == "dl"
+Requires-Dist: tensorflow>=2.14; extra == "dl"
+Provides-Extra: viz
+Requires-Dist: matplotlib>=3.7; extra == "viz"
+Requires-Dist: seaborn>=0.12; extra == "viz"
+Provides-Extra: excel
+Requires-Dist: openpyxl>=3.1; extra == "excel"
+Requires-Dist: xlrd>=2.0; extra == "excel"
+Provides-Extra: all
+Requires-Dist: numba>=0.57; extra == "all"
+Requires-Dist: pyarrow>=12; extra == "all"
+Requires-Dist: pytest; extra == "all"
+Requires-Dist: sqlalchemy>=2.0; extra == "all"
+Requires-Dist: duckdb>=0.9; extra == "all"
+Requires-Dist: scikit-learn>=1.3; extra == "all"
+Requires-Dist: scipy>=1.10; extra == "all"
+Requires-Dist: matplotlib>=3.7; extra == "all"
+Requires-Dist: openpyxl>=3.1; extra == "all"
+Dynamic: license-file
+
+# bandas ⚡ — pandas-compatible, faster + Universal Data Platform
+
+> أُعيدت تسمية المكتبة من `fandas` إلى `bandas` — المستودع الرسمي الآن: https://github.com/salim-studio/bandas
+> للترحيل: `pip uninstall fandas` ثم `pip install bandas-df`، واستبدل `import fandas` بـ `import bandas`.
+
+`bandas` واجهة مطابقة لـ `pandas` (نفس الأسماء: `DataFrame`, `Series`, `read_csv`, `concat`, `merge`...) لكن **أسرع**، ومع **منظومة متكاملة** لقواعد البيانات، السلاسل الزمنية، تجهيز البيانات، تعلم الآلة، التعلم العميق، والتصوير — لتصبح خيار المطورين ومحللي البيانات وعلماء البيانات.
+
+## لماذا bandas؟
+
+1. **تخزين عمودي خالص بـ NumPy** بدون Overhead الـ BlockManager.
+2. **Fast-paths رقمية**: `sum/mean/min/max` عبر numpy مباشرة + `bincount` للـ groupby.
+3. **تسريع اختياري بـ Numba** + **قراءة CSV بـ PyArrow** (أسرع 3-5x).
+4. **SQL حقيقي**: `df.sql("SELECT ...")` بدون أي اعتماد إضافي (sqlite مدمج)، وموصل موحد `Database` لـ sqlite / duckdb / postgres / mysql / sqlalchemy.
+5. **ML/DL جاهز**: `train_test_split`, `KFold`, مقاييس، `StandardScaler`, `OneHotEncoder`, `to_torch`, `to_tensorflow`, `torch_dataset`.
+6. **تحليل متقدم**: `rolling/expanding/ewm`, `shift/diff/pct_change`, `pivot_table/melt/crosstab`, `str/dt accessors`, `LazyFrame` للملفات الضخمة خارج الذاكرة.
+
+## تنصيب
+
+تنصيب واحد يثبّت كل شيء (سرعة + قواعد بيانات + تعلم آلة + تصوير + إكسل):
+
+```bash
+pip install bandas-df
+```
+
+من المصدر:
+
+```bash
+pip install -e .
+```
+
+إضافي فقط للتعلم العميق (torch و tensorflow ضخمان — جيجابايتات — وغير متوفرين لكل نسخ Python):
+
+```bash
+pip install "bandas-df[dl]"
+# أو من المصدر:
+pip install -e ".[dl]"
+```
+
+## 1) أساسيات (أسرع من pandas)
+
+```python
+import bandas as bd
+
+df = bd.DataFrame({"a": [1,2,3], "b": [4.0,5.0,6.0]})
+print(df.head())
+print(df["a"].sum())
+
+df2 = bd.read_csv("data.csv")          # pyarrow إن وجد
+g = df2.groupby("city")["price"].mean()
+print(g.to_pandas())
+
+pdf = df.to_pandas()
+df3 = bd.from_pandas(pdf)
+```
+
+النتيجة على 500k صف: `sum 1.9x`, `mean 1.5x`, `groupby 1.3x` أسرع من pandas (انظر `bench/bench.py`).
+
+## 2) قواعد البيانات / SQL
+
+```python
+# SQL فوري بدون إعداد — يعمل دائماً (sqlite مدمج)
+top = df.sql("SELECT city, AVG(price) AS m FROM df GROUP BY city ORDER BY m DESC")
+
+# استعلام عبر عدة جداول
+from bandas import sql_query
+out = sql_query("SELECT * FROM a JOIN b ON a.k = b.k WHERE a.v > 10", {"a": a, "b": b})
+
+# موصل موحد
+from bandas.sql import Database
+db = Database.sqlite("app.db")          # أو :memory:
+# db = Database.duckdb("data.duckdb")   # يحتاج duckdb
+# db = Database.postgres(user=..., password=..., host=..., db=...)
+db.write(df, "sales")
+print(db.query("SELECT COUNT(*) AS n FROM sales").to_pandas())
+print(db.tables())
+
+# pandas API
+df.to_sql("sales", "sqlite:///app.db")
+df2 = bd.read_sql("SELECT * FROM sales", "sqlite:///app.db")
+```
+
+## 3) سلاسل زمنية ونوافذ
+
+```python
+s = df["price"]
+s.rolling(7).mean()        # متوسط متحرك
+s.expanding().sum()
+s.ewm(span=10).mean()
+s.shift(1); s.diff(); s.pct_change(); s.cumsum()
+df.rolling(30).mean(); df.ewm(alpha=0.3).mean()
+
+bd.to_datetime(["2024-01-01", "2024-02-01"])
+bd.date_range("2024-01-01", periods=12, freq="ME")
+df["d"].dt.year / .month / .day / .dayofweek
+```
+
+## 4) نصوص وفئات وإعادة تشكيل
+
+```python
+df["name"].str.lower().str.strip().str.contains("ali")
+df["name"].str.replace("a", "@").str.split(" ", expand=True)
+bd.get_dummies(df, columns=["city"])
+bd.melt(df, id_vars=["city"])
+bd.pivot_table(df, values="price", index="city", columns="year", aggfunc="mean")
+bd.crosstab(df["city"], df["year"])
+bd.cut(df["age"], bins=[0,18,60,100])
+bd.qcut(df["price"], 4)
+```
+
+## 5) تجهيز البيانات (Preprocessing)
+
+```python
+bd.SimpleImputer(strategy="mean").fit_transform(df)
+bd.StandardScaler().fit_transform(df)
+bd.MinMaxScaler().fit_transform(df)
+bd.RobustScaler().fit_transform(df)
+bd.LabelEncoder().fit_transform(df["city"])
+bd.OneHotEncoder(columns=["city"]).fit_transform(df)
+```
+
+## 6) تعلم الآلة
+
+```python
+X_train, X_test, y_train, y_test = bd.train_test_split(X, y, test_size=0.2, random_state=0)
+list(bd.KFold(n_splits=5).split(X))
+bd.accuracy_score(y, pred); bd.f1_score(y, pred)
+bd.mean_squared_error(y, pred); bd.r2_score(y, pred)
+
+# يعمل مباشرة مع sklearn
+from sklearn.ensemble import RandomForestClassifier
+X, y = bd.to_sklearn_Xy if False else (df.drop(columns=["target"]).to_numpy(), df["target"].to_numpy())
+clf = RandomForestClassifier().fit(X_train_numpy, y_train_numpy)
+```
+
+## 7) التعلم العميق
+
+```python
+X_t, y_t = df.to_torch(target="label")          # torch.Tensor
+loader = bd.dl.torch_dataset(df, target="label", batch_size=64)
+ds = bd.dl.tf_dataset(df, target="label")       # tf.data.Dataset
+```
+
+## 8) ملفات ضخمة (Lazy / out-of-core)
+
+```python
+lf = bd.read_csv_chunked("big.csv", chunksize=100_000)
+out = lf.filter(lambda d: d[d["x"] > 0]).select(["x","y"]).collect()
+print(lf.sum("x"), lf.mean("x"), lf.count())
+```
+
+## 9) تصوير سريع
+
+```python
+df.plot(kind="line", x="date", y="price")
+df.plot(kind="bar", x="city", y="sales")
+df["price"].hist(bins=50)
+bd.viz.scatter(df, "a", "b"); bd.viz.heatmap_corr(df)
+```
+
+## 10) IO شامل
+
+`read_csv / read_parquet / read_json / read_excel / read_html / read_feather / read_orc / read_sql` +
+`to_csv / to_parquet / to_json / to_excel / to_sql` — كلها بنفس توقيع pandas.
+
+## هيكل المشروع
+
+```
+bandas/
+  __init__.py      # الواجهة العامة + concat/merge/options/show_versions
+  series.py        # Series + str/dt + rolling/ewm + cum*/shift/diff + ML helpers
+  dataframe.py     # DataFrame + SQL + stats + reshape + viz + torch
+  groupby.py       # GroupBy السريع + std/var/median/first/last/nunique
+  io.py            # كل قارئات الملفات + chunked
+  sql.py           # sql_query + read_sql/to_sql + Database الموحد
+  reshape.py       # get_dummies/melt/pivot/crosstab/cut/qcut
+  window.py        # Rolling/Expanding/EWM
+  strings.py       # Series.str
+  datetimes.py     # Series.dt + to_datetime/date_range
+  preprocessing.py # Imputer/Scaler/Encoder
+  ml.py            # split/CV/metrics/sklearn bridge
+  dl.py            # torch/tensorflow bridges
+  stats.py         # corr/cov/outliers/summary
+  lazy.py          # LazyFrame خارج الذاكرة
+  viz.py           # رسوم سريعة
+  ops.py / utils.py
+tests/test_basic.py + test_advanced.py (20 اختباراً)
+```
+
+## اختبار و Benchmark
+
+```bash
+pytest -q
+python bench/bench.py
+python -c "import bandas; bandas.show_versions()"
+```

bandas_df-0.3.0/README.md

@@ -0,0 +1,195 @@
+# bandas ⚡ — pandas-compatible, faster + Universal Data Platform
+
+> أُعيدت تسمية المكتبة من `fandas` إلى `bandas` — المستودع الرسمي الآن: https://github.com/salim-studio/bandas
+> للترحيل: `pip uninstall fandas` ثم `pip install bandas-df`، واستبدل `import fandas` بـ `import bandas`.
+
+`bandas` واجهة مطابقة لـ `pandas` (نفس الأسماء: `DataFrame`, `Series`, `read_csv`, `concat`, `merge`...) لكن **أسرع**، ومع **منظومة متكاملة** لقواعد البيانات، السلاسل الزمنية، تجهيز البيانات، تعلم الآلة، التعلم العميق، والتصوير — لتصبح خيار المطورين ومحللي البيانات وعلماء البيانات.
+
+## لماذا bandas؟
+
+1. **تخزين عمودي خالص بـ NumPy** بدون Overhead الـ BlockManager.
+2. **Fast-paths رقمية**: `sum/mean/min/max` عبر numpy مباشرة + `bincount` للـ groupby.
+3. **تسريع اختياري بـ Numba** + **قراءة CSV بـ PyArrow** (أسرع 3-5x).
+4. **SQL حقيقي**: `df.sql("SELECT ...")` بدون أي اعتماد إضافي (sqlite مدمج)، وموصل موحد `Database` لـ sqlite / duckdb / postgres / mysql / sqlalchemy.
+5. **ML/DL جاهز**: `train_test_split`, `KFold`, مقاييس، `StandardScaler`, `OneHotEncoder`, `to_torch`, `to_tensorflow`, `torch_dataset`.
+6. **تحليل متقدم**: `rolling/expanding/ewm`, `shift/diff/pct_change`, `pivot_table/melt/crosstab`, `str/dt accessors`, `LazyFrame` للملفات الضخمة خارج الذاكرة.
+
+## تنصيب
+
+تنصيب واحد يثبّت كل شيء (سرعة + قواعد بيانات + تعلم آلة + تصوير + إكسل):
+
+```bash
+pip install bandas-df
+```
+
+من المصدر:
+
+```bash
+pip install -e .
+```
+
+إضافي فقط للتعلم العميق (torch و tensorflow ضخمان — جيجابايتات — وغير متوفرين لكل نسخ Python):
+
+```bash
+pip install "bandas-df[dl]"
+# أو من المصدر:
+pip install -e ".[dl]"
+```
+
+## 1) أساسيات (أسرع من pandas)
+
+```python
+import bandas as bd
+
+df = bd.DataFrame({"a": [1,2,3], "b": [4.0,5.0,6.0]})
+print(df.head())
+print(df["a"].sum())
+
+df2 = bd.read_csv("data.csv")          # pyarrow إن وجد
+g = df2.groupby("city")["price"].mean()
+print(g.to_pandas())
+
+pdf = df.to_pandas()
+df3 = bd.from_pandas(pdf)
+```
+
+النتيجة على 500k صف: `sum 1.9x`, `mean 1.5x`, `groupby 1.3x` أسرع من pandas (انظر `bench/bench.py`).
+
+## 2) قواعد البيانات / SQL
+
+```python
+# SQL فوري بدون إعداد — يعمل دائماً (sqlite مدمج)
+top = df.sql("SELECT city, AVG(price) AS m FROM df GROUP BY city ORDER BY m DESC")
+
+# استعلام عبر عدة جداول
+from bandas import sql_query
+out = sql_query("SELECT * FROM a JOIN b ON a.k = b.k WHERE a.v > 10", {"a": a, "b": b})
+
+# موصل موحد
+from bandas.sql import Database
+db = Database.sqlite("app.db")          # أو :memory:
+# db = Database.duckdb("data.duckdb")   # يحتاج duckdb
+# db = Database.postgres(user=..., password=..., host=..., db=...)
+db.write(df, "sales")
+print(db.query("SELECT COUNT(*) AS n FROM sales").to_pandas())
+print(db.tables())
+
+# pandas API
+df.to_sql("sales", "sqlite:///app.db")
+df2 = bd.read_sql("SELECT * FROM sales", "sqlite:///app.db")
+```
+
+## 3) سلاسل زمنية ونوافذ
+
+```python
+s = df["price"]
+s.rolling(7).mean()        # متوسط متحرك
+s.expanding().sum()
+s.ewm(span=10).mean()
+s.shift(1); s.diff(); s.pct_change(); s.cumsum()
+df.rolling(30).mean(); df.ewm(alpha=0.3).mean()
+
+bd.to_datetime(["2024-01-01", "2024-02-01"])
+bd.date_range("2024-01-01", periods=12, freq="ME")
+df["d"].dt.year / .month / .day / .dayofweek
+```
+
+## 4) نصوص وفئات وإعادة تشكيل
+
+```python
+df["name"].str.lower().str.strip().str.contains("ali")
+df["name"].str.replace("a", "@").str.split(" ", expand=True)
+bd.get_dummies(df, columns=["city"])
+bd.melt(df, id_vars=["city"])
+bd.pivot_table(df, values="price", index="city", columns="year", aggfunc="mean")
+bd.crosstab(df["city"], df["year"])
+bd.cut(df["age"], bins=[0,18,60,100])
+bd.qcut(df["price"], 4)
+```
+
+## 5) تجهيز البيانات (Preprocessing)
+
+```python
+bd.SimpleImputer(strategy="mean").fit_transform(df)
+bd.StandardScaler().fit_transform(df)
+bd.MinMaxScaler().fit_transform(df)
+bd.RobustScaler().fit_transform(df)
+bd.LabelEncoder().fit_transform(df["city"])
+bd.OneHotEncoder(columns=["city"]).fit_transform(df)
+```
+
+## 6) تعلم الآلة
+
+```python
+X_train, X_test, y_train, y_test = bd.train_test_split(X, y, test_size=0.2, random_state=0)
+list(bd.KFold(n_splits=5).split(X))
+bd.accuracy_score(y, pred); bd.f1_score(y, pred)
+bd.mean_squared_error(y, pred); bd.r2_score(y, pred)
+
+# يعمل مباشرة مع sklearn
+from sklearn.ensemble import RandomForestClassifier
+X, y = bd.to_sklearn_Xy if False else (df.drop(columns=["target"]).to_numpy(), df["target"].to_numpy())
+clf = RandomForestClassifier().fit(X_train_numpy, y_train_numpy)
+```
+
+## 7) التعلم العميق
+
+```python
+X_t, y_t = df.to_torch(target="label")          # torch.Tensor
+loader = bd.dl.torch_dataset(df, target="label", batch_size=64)
+ds = bd.dl.tf_dataset(df, target="label")       # tf.data.Dataset
+```
+
+## 8) ملفات ضخمة (Lazy / out-of-core)
+
+```python
+lf = bd.read_csv_chunked("big.csv", chunksize=100_000)
+out = lf.filter(lambda d: d[d["x"] > 0]).select(["x","y"]).collect()
+print(lf.sum("x"), lf.mean("x"), lf.count())
+```
+
+## 9) تصوير سريع
+
+```python
+df.plot(kind="line", x="date", y="price")
+df.plot(kind="bar", x="city", y="sales")
+df["price"].hist(bins=50)
+bd.viz.scatter(df, "a", "b"); bd.viz.heatmap_corr(df)
+```
+
+## 10) IO شامل
+
+`read_csv / read_parquet / read_json / read_excel / read_html / read_feather / read_orc / read_sql` +
+`to_csv / to_parquet / to_json / to_excel / to_sql` — كلها بنفس توقيع pandas.
+
+## هيكل المشروع
+
+```
+bandas/
+  __init__.py      # الواجهة العامة + concat/merge/options/show_versions
+  series.py        # Series + str/dt + rolling/ewm + cum*/shift/diff + ML helpers
+  dataframe.py     # DataFrame + SQL + stats + reshape + viz + torch
+  groupby.py       # GroupBy السريع + std/var/median/first/last/nunique
+  io.py            # كل قارئات الملفات + chunked
+  sql.py           # sql_query + read_sql/to_sql + Database الموحد
+  reshape.py       # get_dummies/melt/pivot/crosstab/cut/qcut
+  window.py        # Rolling/Expanding/EWM
+  strings.py       # Series.str
+  datetimes.py     # Series.dt + to_datetime/date_range
+  preprocessing.py # Imputer/Scaler/Encoder
+  ml.py            # split/CV/metrics/sklearn bridge
+  dl.py            # torch/tensorflow bridges
+  stats.py         # corr/cov/outliers/summary
+  lazy.py          # LazyFrame خارج الذاكرة
+  viz.py           # رسوم سريعة
+  ops.py / utils.py
+tests/test_basic.py + test_advanced.py (20 اختباراً)
+```
+
+## اختبار و Benchmark
+
+```bash
+pytest -q
+python bench/bench.py
+python -c "import bandas; bandas.show_versions()"
+```

bandas_df-0.3.0/bandas/__init__.py

@@ -0,0 +1,120 @@
+"""bandas — pandas-compatible, faster DataFrame library + universal data platform.
+
+Core (fast NumPy paths) + SQL/DB + time-series + preprocessing + ML/DL + viz + lazy.
+"""
+from __future__ import annotations
+
+from .series import Series
+from .dataframe import DataFrame
+from .io import (
+    read_csv, read_parquet, read_json, read_excel, read_html,
+    read_feather, read_orc, read_sql, read_csv_chunked,
+    from_pandas, to_pandas,
+)
+from .sql import sql_query, to_sql, Database
+from .reshape import get_dummies, melt, pivot_table, crosstab, cut, qcut
+from .datetimes import to_datetime, date_range
+from .lazy import LazyFrame
+from . import preprocessing, ml, stats, viz, dl
+from .preprocessing import SimpleImputer, StandardScaler, MinMaxScaler, RobustScaler, LabelEncoder, OneHotEncoder
+from .ml import train_test_split, KFold, accuracy_score, mean_squared_error, mean_absolute_error, r2_score, precision_score, recall_score, f1_score
+
+__version__ = "0.3.0"
+__all__ = [
+    "DataFrame", "Series", "Database", "LazyFrame",
+    "read_csv", "read_parquet", "read_json", "read_excel", "read_html",
+    "read_feather", "read_orc", "read_sql", "read_csv_chunked",
+    "from_pandas", "to_pandas", "sql_query", "to_sql",
+    "concat", "merge", "merge_ordered", "merge_asof",
+    "get_dummies", "melt", "pivot_table", "crosstab", "cut", "qcut",
+    "to_datetime", "date_range",
+    "SimpleImputer", "StandardScaler", "MinMaxScaler", "RobustScaler",
+    "LabelEncoder", "OneHotEncoder",
+    "train_test_split", "KFold",
+    "accuracy_score", "mean_squared_error", "mean_absolute_error", "r2_score",
+    "precision_score", "recall_score", "f1_score",
+    "preprocessing", "ml", "stats", "viz", "dl",
+    "show_versions", "options",
+]
+
+
+def concat(objs, axis=0, ignore_index=False):
+    import numpy as np
+    from .series import Series as _S
+    if not objs:
+        return DataFrame({})
+    if isinstance(objs[0], _S):
+        arr = np.concatenate([np.asarray(o._data) for o in objs])
+        idx = np.concatenate([np.asarray(o._index) for o in objs]) if not ignore_index else np.arange(len(arr))
+        return _S(arr, index=idx, name=objs[0]._name)
+    if axis == 1:
+        out_cols, out_data = [], {}
+        n = max(len(o) for o in objs)
+        for o in objs:
+            for c in o._columns:
+                name = c
+                k = 0
+                while name in out_data:
+                    k += 1
+                    name = f"{c}_{k}"
+                out_cols.append(name)
+                a = np.asarray(o._cols[c])
+                if len(a) < n:
+                    pad = np.full(n - len(a), np.nan, dtype=np.float64 if a.dtype.kind in "iuf" else object)
+                    a = np.concatenate([a.astype(pad.dtype) if a.dtype != pad.dtype else a, pad])
+                out_data[name] = a
+        nd = DataFrame.__new__(DataFrame)
+        nd._columns = out_cols
+        nd._cols = out_data
+        nd._index = np.arange(n) if ignore_index else np.asarray(objs[0]._index)
+        return nd
+    cols = list(objs[0]._columns)
+    for o in objs[1:]:
+        for c in o._columns:
+            if c not in cols:
+                cols.append(c)
+    out = {c: [] for c in cols}
+    idx_parts = []
+    for o in objs:
+        for c in cols:
+            if c in o._cols:
+                out[c].append(np.asarray(o._cols[c]))
+            else:
+                out[c].append(np.full(len(o), np.nan))
+        idx_parts.append(np.asarray(o._index))
+    data = {c: np.concatenate(v) if v else np.array([]) for c, v in out.items()}
+    idx = np.arange(sum(len(o) for o in objs)) if ignore_index else np.concatenate(idx_parts) if idx_parts else np.array([])
+    return DataFrame(data, index=idx)
+
+
+def merge(left, right, on=None, how="inner", left_on=None, right_on=None):
+    return left.merge(right, on=on, how=how, left_on=left_on, right_on=right_on)
+
+
+def merge_ordered(left, right, on=None, how="outer", **kw):
+    out = merge(left, right, on=on, how=how, **kw)
+    return out.sort_values(on) if on else out
+
+
+def merge_asof(left, right, on=None, direction="backward"):
+    import pandas as pd
+    return DataFrame(pd.merge_asof(left.to_pandas().sort_values(on), right.to_pandas().sort_values(on), on=on, direction=direction))
+
+
+class _Options(dict):
+    def __repr__(self): return f"bandas.options({dict(self)})"
+
+options = _Options({"display.max_rows": 10, "compute.numba": True, "io.pyarrow": True})
+
+
+def show_versions():
+    import sys, numpy, pandas
+    info = {"python": sys.version.split()[0], "bandas": __version__, "numpy": numpy.__version__, "pandas": pandas.__version__}
+    for mod in ("pyarrow", "numba", "sqlalchemy", "duckdb", "sklearn", "torch", "tensorflow", "matplotlib"):
+        try:
+            m = __import__(mod)
+            info[mod] = getattr(m, "__version__", "installed")
+        except Exception:
+            info[mod] = "not installed"
+    for k, v in info.items(): print(f"{k}: {v}")
+    return info