fix #388

qiita_db/metadata_template/prep_template.py

@@ -13,6 +13,7 @@
 from time import strftime
 from copy import deepcopy
 import warnings
+from skbio.util import find_duplicates
 
 import pandas as pd
 
@@ -23,6 +24,30 @@
 from .base_metadata_template import BaseSample, MetadataTemplate
 
 
+def _check_duplicated_columns(prep_cols, sample_cols):
+    r"""Check for duplicated colums in the prep_cols and sample_cols
+
+    Parameters
+    ----------
+    prep_cols : list of str
+        Column names in the prep info file
+    sample_cols : list of str
+        Column names in the sample info file
+
+    Raises
+    ------
+    QiitaDBColumnError
+        If there are duplicated columns names in the sample and the prep
+    """
+    prep_cols.extend(sample_cols)
+    dups = find_duplicates(prep_cols)
+    if dups:
+        raise qdb.exceptions.QiitaDBColumnError(
+            'Duplicated column names in the sample and prep info '
+            'files: %s. You need to delete that duplicated field' %
+            ','.join(dups))
+
+
 class PrepSample(BaseSample):
     r"""Class that models a sample present in a PrepTemplate.
 
@@ -113,6 +138,8 @@ def create(cls, md_template, study, data_type, investigation_type=None):
                 pt_cols.update(PREP_TEMPLATE_COLUMNS_TARGET_GENE)
 
             md_template = cls._clean_validate_template(md_template, study.id)
+            _check_duplicated_columns(list(md_template.columns),
+                                      study.sample_template.categories())
 
             # Insert the metadata template
             sql = """INSERT INTO qiita.prep_template
@@ -358,6 +385,10 @@ def can_be_extended(self, new_samples, new_columns):
                                        "template has already been processed. "
                                        "No new samples can be added to the "
                                        "prep template")
+
+        _check_duplicated_columns(list(new_columns), qdb.study.Study(
+            self.study_id).sample_template.categories())
+
         return True, ""
 
     @property

qiita_db/metadata_template/test/test_prep_template.py

@@ -243,6 +243,11 @@ def test_can_be_extended(self):
                    "template")
         self.assertEqual(obs_msg, exp_msg)
 
+    def test_can_be_extended_duplicated_column(self):
+        """test if the template can be extended"""
+        with self.assertRaises(qdb.exceptions.QiitaDBColumnError):
+            self.prep_template.can_be_extended([], ["season_environment"])
+
     def test_metadata_headers(self):
         PT = qdb.metadata_template.prep_template.PrepTemplate
         obs = PT.metadata_headers()
@@ -1003,6 +1008,13 @@ def test_create_investigation_type_error(self):
                 self.metadata, self.test_study, self.data_type_id,
                 'Not a term')
 
+    def test_create_duplicated_column_error(self):
+        """Create raises an error if the prep has a duplicated column name"""
+        self.metadata['season_environment'] = self.metadata['primer']
+        with self.assertRaises(qdb.exceptions.QiitaDBColumnError):
+            qdb.metadata_template.prep_template.PrepTemplate.create(
+                self.metadata, self.test_study, self.data_type_id)
+
     def test_delete_error(self):
         """Try to delete a prep template that already has preprocessed data"""
         with self.assertRaises(qdb.exceptions.QiitaDBExecutionError):

qiita_db/support_files/patches/45.sql

@@ -0,0 +1,4 @@
+-- Dec 15, 2016
+-- Making sure there are no duplicated columns, much easier via python
+
+SELECT 42;

qiita_db/support_files/patches/python_patches/45.py

@@ -0,0 +1,36 @@
+from future.utils import viewitems
+
+from qiita_db.metadata_template.sample_template import SampleTemplate
+from qiita_db.metadata_template.prep_template import PrepTemplate
+from qiita_db.sql_connection import TRN
+
+with TRN:
+    # a few notes: just getting the preps with duplicated values; ignoring
+    # column 'sample_id' and tables 'study_sample', 'prep_template',
+    # 'prep_template_sample'
+    sql = """SELECT table_name, array_agg(column_name::text)
+                FROM information_schema.columns
+                WHERE column_name IN %s
+                    AND column_name != 'sample_id'
+                    AND table_name LIKE 'prep_%%'
+                    AND table_name NOT IN (
+                        'prep_template', 'prep_template_sample')
+                GROUP BY table_name"""
+    # note that we are looking for those columns with duplicated names in
+    # the headers
+    TRN.add(sql, [tuple(
+        set(PrepTemplate.metadata_headers()) &
+        set(SampleTemplate.metadata_headers()))])
+    overlapping = dict(TRN.execute_fetchindex())
+
+# finding actual duplicates
+for table_name, cols in viewitems(overlapping):
+    # leaving print so when we patch in the main system we know that
+    # nothing was renamed or deal with that
+    print table_name
+    with TRN:
+        for c in cols:
+            sql = 'ALTER TABLE qiita.%s RENAME COLUMN %s TO %s_renamed' % (
+                table_name, c, c)
+            TRN.add(sql)
+        TRN.execute()

qiita_db/test/test_commands.py

@@ -457,7 +457,7 @@ def test_update_artifact_from_cmd(self):
 
 PREP_TEMPLATE = (
     'sample_name\tbarcode\tcenter_name\tcenter_project_name\t'
-    'description\tebi_submission_accession\temp_status\tprimer\t'
+    'description_prep\tebi_submission_accession\temp_status\tprimer\t'
     'run_prefix\tstr_column\tplatform\tlibrary_construction_protocol\t'
     'experiment_design_description\tinstrument_model\n'
     'SKB7.640196\tCCTCTGAGAGCT\tANL\tTest Project\tskb7\tNone\tEMP\t'